The Memory Illusion

by Dr Julia Shaw

  ‘War-zone.’

  As it turned out, one of the notorious Bacon Brothers had just been shot dead. The brothers were a trio of gangsters who had been implicated in a rash of homicides in the Greater Vancouver area, along with drug production and trafficking. Jonathan Bacon and his family had just been attacked and gunned down in broad daylight by rival gangsters. And the public had documented everything.

  We are prone to whip out our phones to film, snap, summarise and post things at the first sign of potential import. Never in history have we had such reliable, independent and ample documentation of important historic events. This ability to corroborate our own appraisals of situations is amazing, but it can also lead to memory conformity – when our way of thinking and our memories become an amalgamated version of the accounts we have seen and heard, and it becomes impossible to distinguish what any one person actually witnessed themselves.

  Almost everyone in Kelowna seems to remember the Bacon Brothers shooting the same way. When you talk to people about it their accounts are amazingly, even impossibly, similar. You can probably think of events that you witnessed or were peripherally involved in where the same thing is true. Educational researcher Brian Clark from Western Illinois University argues in his 2013 article cheekily entitled ‘From yearbooks to Facebook’9 that such effects may be because our memory has gone through a transition due to the internet and social media – ‘the distinction between public memory and private memory … has been blurred to the point of erasure.’

  Research has investigated memory conformity in various settings, particularly in eyewitness accounts. In a 2003 paper Fiona Gabbert, Amina Memon and Kevin Allan from the University of Aberdeen looked at how eyewitnesses can influence each other.10 To test this, they asked two groups of participants to separately watch a video of an event. All participants saw a 90-second video showing a woman going into an empty university office to return a book. Unbeknown to them, there were two different versions of the video, shot from different angles. This meant that participants ended up with one of two sets of information about the event.

  The researchers themselves described the difference as follows: ‘from perspective “A” (but not perspective B) it is possible to read the title of the book that the girl is carrying, and also observe that she throws a [paper] note into a dustbin when leaving the room. From perspective “B” (but not perspective A) the girl is seen checking the time on her watch, as well as committing an opportunistic crime (sliding a £10 [bank]note out of a wallet and putting it into her own pocket).’

  Half of the participants were then asked to work together as teams of two to complete a questionnaire about what happened, while the other half completed the questionnaire alone. Then, after a 45-minute interval, all participants were asked individually about what happened. The researchers found that 71 per cent of those in the co-witness condition reported knowing details that they had obtained through discussing the event with their partners. Further, 60 per cent of those who watched the video from perspective A, where they could not actually see the opportunistic crime taking place, reported that the girl in the video was guilty of a crime. Those in the co-witness group had included on average 21 details that they stole from the other witness. As could be expected, people who completed the questionnaire alone only reported details from the video they had actually seen. Participants in the co-witness group had severely embellished their memory reports with details they had not actually witnessed themselves.

  Research like this explores what is referred to as ‘post-event information’ – information that can influence our memories if we encounter it after we experience or witness an event. It might come from many possible sources – discussing the event with others in person or online, reading articles about the event or related events, seeing photos taken by ourselves or others, to name but a few. Any source of information has the potential to change our memories post hoc.

  According to psychological scientist Alan Brown11 from Southern Methodist University and his colleagues, another source of false memories is memory borrowing, where someone directly appropriates someone else’s autobiographical memory and relates it as their own. Brown and his team published a paper in 2015 investigating this phenomenon and found that of 447 students who participated in a survey on the topic, 47 per cent answered yes to the question ‘Have you ever heard someone’s personal experience and later told it to others as if it happened to you?’ This means that the students had knowingly claimed authorship of another’s autobiographical memory, at least temporarily. Although it may be done consciously, this kind of temporary borrowing can lead to later memory attribution issues, as 27 per cent of the participants also claimed that they had memories which could be their own but could be borrowed from someone else’s account of an event – and they were unsure which was the case.

  Brown’s research also demonstrated that sometimes memory thieves get caught; 53 per cent of participants claimed that they had heard someone tell one of their stories as if it was their own, and 57 per cent claimed that they had disagreed with someone over whether an incident had happened to themselves or to the other person. I personally find this kind of memory thievery particularly applies to family stories, where I occasionally catch myself having to seek confirmation of what really happened from another family member.

  So it’s clear that memories are contagious; if I let out one of my memories, it is possible for you to catch it and make it your own. And as we merge details from other sources into our own retellings of an event, we have the potential to incorporate both accurate and inaccurate details. In a paper published in 2001, Henry Roediger and colleagues from Washington University coined a good term for this: the social contagion of memory. They showed that one person’s memory can be influenced by another’s memory errors. A sort of false memory proliferation effect. But why we are so prone to this effect? Researchers argue that it is due to two factors. The first is basic memory distortion; if another person tells you their version of an event, your brain may make new connections that subsequently interfere with your own original memory of it. This is in line with the misinformation and imagination inflation research that we discussed in earlier chapters. The second is source confusion, where we forget the source of information we remember, which can lead us to assume we experienced things that were only told to us.

  According to Brown and the other authors of the memory thief experiment, social influences come into play in a number of ways; ‘These behaviors appear primarily motivated by a desire to permanently incorporate others’ experiences into one’s own autobiographical record (appropriation), but other reasons include to temporarily create a more coherent or engaging conversational exchange (social connection), simplify conveying somebody else’s interesting experience (convenience), or make oneself look good (status enhancement).’ These seem like reasons that are positive and often intentional. But there are scientists who strongly argue for yet another possible social influence: conformity.

  Groupiness

  The classic studies that first showed our conformity to information when it is provided by others were conducted by psychological scientist Solomon Asch from Swarthmore College in 1956.12 He found that if we ask people in a group to judge whether or not two lines on a paper are a similar length, their answers will change depending on what others in the group say. Solomon looked at this by planting a number of research confederates who were instructed to give an obviously wrong answer in the room with the participants. The participants thought that all members of the group were other participants, and did not know that they were the only ones who were actually being studied. Indeed, people are often willing to provide a very obviously incorrect answer if it conforms with the answer that everyone else is giving. While we might be comfortable accepting that some people are naturally ‘followers’ and will inevitably behave that way, what is shocking is that nearly 75 per cent of the participants in the experiments conformed to the obviously incorrect answer provided by the
group at least once, demonstrating that actually the majority of us can be influenced by our peers. We can all be victims of situational demands.

  When asked later why they had conformed, most participants claimed that they knew the answer was wrong but did not want to stand out. Some, however, claimed that they had really believed that the group must have known the answer better than they did. In 1955 social scientists Morton Deutsch and Harold Gerard from New York University13 went on to classify social influences of this kind as either normative or informational.

  Normative influences are the influences of groups on their members – situations where we do not want to stand out, regardless of whether we believe the group to be correct or not. Informational social influences are also facilitated by groups, but do not necessarily require them. They are instances in which we believe that another person is better informed than us, so we adopt their information on the basis that it is probably correct – a situation where a group or, say, an interviewer really does know the correct answer.

  These influences help to explain why one person might adopt the account of another. They either might not want to upset the other person by disagreeing (a normative influence), or they may genuinely believe that the other person has a better memory of the account than they do (an informational influence). Of course these social influences are not always a bad thing. If a group of people are running, perhaps they know that there is a fire and you don’t – conformity can save lives. It also certainly has benefits for easing conversation and collaborations between members of a group if our memories conform as well. But these social influences become a problem when it is post-event misinformation that is being spread by them, causing erroneous details to become woven into our memories in ways that can never be untangled.

  But that’s not all. Deutsch and Gerard coined the term ‘groupiness’ to describe how cohesive a particular group is – how much its members tend to conform. In sociology the term for this is entitativity, essentially meaning the degree to which the group works as a single entity. We tend to divide the world into ‘in-groups’ and ‘out-groups’, meaning those that we identify ourselves as being a part of, and everyone else. For example, your in-group may be your university alma mater, while the out-group may include students from a rival institution.

  Dan Ariely, professor of psychology and behavioural economics at Duke University and the author of the bestselling Predictably Irrational,14 argues that our membership of groups makes us exactly that: predictably irrational. Ariely and his colleagues15 have conducted numerous experiments which have demonstrated that when members of our in-group do something, we are likely to follow suit. This is true for the good and the bad – for example, we are more likely to cheat if at least one of our in-group members does as well. Ariely’s research has also demonstrated that we are less likely to conform to those with whom we do not identify – out-group members. This is presumably a deliberate delineation from the behaviour of our rivals – we are not like them – as well as an implicit display of solidarity with our in-group brethren – look how we share values.

  Bearing in mind all these influencing factors, many researchers argue that we need to keep witnesses in police proceedings separate from one another, to avoid tainting effects. And police need to understand that consistency between reports may not necessarily indicate accuracy, only conformity.

  Further, the advent of social media has enormously multiplied the potential sources of social influence and misinformation – a friend’s Facebook update, a Twitter post by a stranger, a Reddit discussion thread. It seems as though we no longer have full ownership of events in our lives, and are instead living in a time of intense ‘transactive memory’, as memory researcher Daniel Wegner16 from the University of Virginia would say. Transactive memories, like our online interactions, are memories that are collectively formed, updated, and perhaps most importantly of all, stored.

  Digital amnesia

  According to a powerful article entitled ‘Google effects on memory’ by psychological researcher Betsy Sparrow from Columbia University and her colleagues, ‘The Internet has become a primary form of external or transactive memory, where information is stored collectively outside ourselves.’

  Sparrow and her team conducted a series of four studies to investigate the consequences of having information constantly at our fingertips. In the first she asked participants to answer a series of tricky trivia questions. She then had them do a word-sorting task that measured the speed at which they classified computer-related words and other words. She found that participants who encountered questions to which they did not know the answers were much faster at sorting computer-related words. She took this to indicate that these participants, when faced with questions to which they did not know the answers, had been thinking about search-engine-related words such as Google and Yahoo. This was taken to indicate that our minds quite automatically wander to search engines as ways of answering informational questions. In other words, when we encounter facts we don’t know, we automatically think ‘I should Google it.’

  The second experiment Sparrow did turned the same trivia questions into statements. For example, she would present the fact ‘An ostrich’s eye is bigger than its brain.’ The participants would then type out the fact on a computer, to make sure they were paying attention. Half of her participants were told that the facts they had typed would be saved for later, while the other half were told that the facts would not be saved. Afterwards, participants were asked to write down as many facts as they could remember. Those who were told that the information would be saved performed worse when their memories were tested than those who did not expect the information to be saved. Sparrow and her colleagues claim that this ties in with the idea that knowing we can always access information later on makes us less likely to put in the effort to try to remember it, therefore decreasing our actual memory for it subsequently.

  She repeated the method of study two in study three, except that this time she told participants that the information would either be saved in a specific spot, saved in general, or erased; and instead of having the participants write down the trivia, they did a recognition task. Participants were shown all 30 trivia questions again, but half of them had been slightly altered; the other half were exactly the same as they had seen previously. Participants had to judge whether or not the statements were exactly as they had seen them or not. Again, those who had been told the information they had typed would be erased correctly recognised more of the facts. It seems that we really are less likely to remember information if we think it will be available to us later in a digital form, a phenomenon that is sometimes referred to as digital amnesia. In an age where information is almost always available to us later, this can have profound implications for how we remember it.

  Finally, in study four, Sparrow found something particularly peculiar. This time participants were told that all statements would be saved and accessible in one of six folders. For example, participants were told that the fact ‘The space shuttle Columbia disintegrated during re-entry over Texas in Feb. 2003’ had been saved into a folder named FACTS, DATA, INFO, NAMES, ITEMS or POINTS. When tested on their ability to remember and write down the facts later, it was found that they were more likely to remember where the statements were saved rather than the content of the statements themselves. On top of this preference, participants were particularly bad at recalling both the statement and the folder it was saved in. So if they remembered the statement they were unlikely to remember its location, but if they did not remember the statement they were highly likely to remember its location.

  It appears that our brains are cognitive misers, picking whichever information appears easier to remember – either the statement or the location where it could be found again. As Sparrow puts it: ‘We are becoming symbiotic with our computer tools, growing into interconnected systems that remember less by knowing information than by knowing where the information can be found.’ Take phone numbers as an
example. According to internet security company Kapersky Lab, 50 per cent of people cannot remember the phone number of their partner, and 71 per cent cannot remember the phone numbers of their own children – but I bet all of them know where they can find these numbers on their phones.17

  Outsourcing our information storage in this way presumably means we are potentially even more vulnerable to the kind of post-event misinformation effects mentioned earlier. However, it can free up our cognitive resources to remember other things to which we are less likely to have immediate access elsewhere. We can always look up the name and the fact later, as long as we remember the gist of the information we want to find. Understanding this impact the digital age has had on the way we handle information has the potential to change our approach to education quite radically.

  ‘Perhaps those who teach in any context, be they college professors, doctors or business leaders, will become increasingly focused on imparting greater understanding of ideas and ways of thinking, and less focused on memorisation,’ Sparrow conjectures. Being less focused on imparting specific detailed information that students could easily find online, we could instead teach critical thinking so when people inevitably do Google it, they at least know how to find high-quality information and analyse it. Beyond the fact that we seem to encode and remember information differently depending on whether we can access it again later, there are other ways our growing reliance on media can alter the quality of our memories.

  You’re uglier than you think

  Did you know that strangers know what you look like better than you do? Oh, and you also aren’t as attractive as you think. I know, it would have been kinder not to tell you. The reason for this is twofold, down to both basic memory processes and the way we use technology.