Science Focus - the home of BBC Science Focus Magazine
As far as I can remember © Getty Images

As far as I can remember

Published: 01st January, 2009 at 00:00
Subscribe to BBC Science Focus Magazine and get 6 issues for just £9.99

Even in these days of DNA tests and other forensic techniques, witness testimony still plays a pivotal role in court cases. But how reliable are our memories? Andy Ridgway finds we know less than we think.

Check your facts before giving evidence – your memory can deceive you in the strangest ways.

Most of us have some recollection of the 2005 terrorist attacks in London. It could well be a mental image of a red double-decker bus in Tavistock Square with its roof ripped off by the force of the blast. That’s not surprising given the number of photographs of the stricken bus that were carried in newspapers in the days after the attack.

Adults have ‘childhood amnesia’ and remember little or nothing of what happened to them aged three and below. It’s often not until aged eight or above that many memories can be recalled.
But what about CCTV footage? Do you remember seeing a video of the bus exploding? What can you see in that video?
Well, the truth is, you shouldn’t be able to see anything in your mind’s eye because such CCTV footage simply doesn’t exist. But don’t worry. If it only took a suggestion that you may have seen a video of the explosion to create a mental image in your mind, you’re not alone. In fact, in a study carried out by Dr James Ost at the University of Portsmouth, 40 per cent of people claimed to have seen this nonexistent footage. Some even went on to describe what happened in vivid detail.
Many of us think we have a good memory. After all, it’s got us through the odd exam. But what Ost’s study clearly demonstrates is just how malleable our memories are. ‘Facts’ from the past can become mangled in our minds. And it can simply be the fact that we’ve been asked about something, such as a nonexistent video clip, that can alter our memory.
In many cases, a dodgy memory is not a problem. It just means we send a birthday card late or a story we tell at a party is slightly distorted. But sometimes the contents of our memories can have huge consequences – landing people behind bars or even, in the US, on Death Row.
In 1998, an American study calculated that in 95 per cent of felony cases – the more serious crimes – witness evidence (in other words, people’s memories) was the only evidence heard in court. In the UK, despite DNA and other forensic evidence being used more regularly, witness memories are still a vital part of court proceedings.
Even before a case gets to court, a few dodgy memories can get an investigation off to a bad start. In the sniper attacks that took place in the Washington DC area in 2002, witnesses reported seeing a white van or truck fleeing several of the crime scenes. A white vehicle may have been seen near one of the first shootings and the media repeated this. When they were caught, the sniper suspects were actually driving a blue car. It seems many witness memories had been contaminated by the media reports.

Careless memories

Among psychologists, the fragility of our memories is well known. But there’s a concern that those involved in the judicial system don’t have a clear understanding of the inner workings of our mind. One of Britain’s leading memory experts, Professor Martin Conway, was so concerned about this that in 2008, he and a handful of other psychologists put together a report to arm police officers, lawyers and judges with the latest scientific knowledge on memory.
Conway says misconceptions abound. “People have this belief that if you remember very perceptual sorts of details, like the colour of the wallpaper or the swirly pattern on the carpet, then your memory is highly likely to be correct. That’s completely the reverse of what we know. There’s another belief that if there are no gaps in your memory, and it’s all very fluent, you are a reliable witness.”
Conway, professor of cognitive psychology at the University of Leeds, has been called in as an expert on memory in many court cases. He says that even immediately after a particularly traumatic event, our memories may be poor. “Often victims of violent crimes have really bashed up memories. They are traumatic experiences and the memories are fragmented; some bits are vivid and some bits are clearly distorted. A smart barrister would say, ‘So you can’t tell me what he was wearing, how can we rely on your claim that he raped you?’”
And while we may not remember details that would seem obvious, there’s no rhyme or reason to what does come to mind. “If you experience a traumatic event, you tend to come out with three to five hotspots – moments of mental imagery that are often visual and destabilise you emotionally,” says Conway. “A good example is in car crashes, some people have a vivid memory of the texture of the material on the inside of the car door because at the moment of impact they looked away, and that’s what they saw.”
Many people believe they have seen video footage of the moment Princess Diana’s car crashed – but no such footage actually exists.
Clearly remembering only fragments of a traumatic event, whether it’s a train crash or even a sexual assault, can present problems. “The police always want to get a coherent story, whereas the truth of the matter is that when you remember something from the past, you remember it in fragments, often out of order,” says Conway. “And ideally, that’s what the police should be attempting to get.”
Police in England and Wales have a five-stage model for interviewing witnesses and suspects called PEACE. This advocates allowing the interviewee to provide, at least in the first instance, a ‘free recall’ of what happened. It’s a technique that’s widely regarded among psychologists as being the best strategy to eliminate the possibility of ‘false memories’.
The danger of false memories is demonstrated by some statistics from the other side of the Atlantic. In 2001, reporters working for the Chicago Tribune found that since 1991, 247 murder cases were thrown out by judges or juries because the statements obtained in police interviews were either completely false or tainted. In some cases, the accused were blatantly not guilty – they were actually behind bars when the murder was committed.
Respected psychologist Professor Chuck Brainerd at Cornell University in the US says asking the wrong kinds of questions can lead to this kind of result. “In a criminal case, like murder, you need motive, method and opportunity,” he says. “If motive and method are nailed down and you give information about your whereabouts that’s wrong because of suggestive interviewing techniques, that in effect is a confession.”
Brainerd says there’s a hierarchy of questions that the police can ask. At the one end, there’s the shining beacon of truth – free recall – when officers are most likely to get accurate information from a suspect or witness’s memory. At the other end of the spectrum, you ask a question with a suggestion: “a knife was used, wasn’t it?” or, “he was driving a red Ford Focus, wasn’t he?”.
Asking for free recall in an interview is a way to tap into people’s ‘verbatim memories’ – a sentence-by-sentence, image-by-image recollection of an event. These are memories that have been found to be stored in the temporal lobes at the side of the brain. Verbatim memories tend to be far more accurate than the other kind we store – ‘gist memories’, which are our interpretation of an experience. If it’s a conversation we’re remembering, for instance, it’s our brief synopsis of it. Our understanding of what was being said. These kind of memories, says Brainerd, are stored in the pre-frontal cortex, the area at the front of the brain that’s known to be involved in our processing of the information we’re presented with. And these gist memories are much more likely to contain false information.
The detailed verbatim memories tend to take a lot more ‘brain power’ to store. So, over time, they tend to be replaced with less costly gist memories – and our memories will contain more and more inaccuracies. Among other things, gist memories always tend towards the ‘norm’ of a situation. So if you always tend to drink red wine at parties, that’s what you’re likely to remember in your longer-term gist memory. This means that speaking to witnesses as soon after the event as possible is vital.

Fuzzy logic

Brainerd and his Cornell colleague Valerie Reyna built their ideas on gist and verbatim memories into what they call their Fuzzy Trace Theory. And while the guidance it gives on how and when to interview witnesses and suspects is all well and good, it’s another of the theory’s predictions that could shake the criminal justice system on both sides of the Atlantic.
“The basic belief of courts in Great Britain, the US and Canada is that children are more prone to false memory problems,” says Brainerd. “So all things being equal, if you have a child who says A happened and an adult who says B happened, you assign more credibility to the adult.” But in 2008, Reyna and Brainerd announced that their research proved the situation is a little more complicated. And in some circumstances, a child’s memories are far more accurate than an adult’s. For court cases that involve children, such as allegations of sexual abuse, the implications are huge.
After carrying out tests on volunteers, they found that if you are dealing with a situation that both adults and children understand, an adult’s memory is better. But when it comes to situations that children don’t fully understand, it’s their memories that have the upper hand. “You get these kinds of situations in the law all the time,” says Brainerd. “It’s ones in which you have some kind of meaning, some kind of concept that is sophisticated and adults are going to grasp but kids aren’t, for example relationships, ethnic and gender stereotypes and sex.”
It seems that in some circumstances, our knowledge of the world and our preconceived ideas mean that we are more likely to store gist memories, whereas children, who don’t summarise what they don’t fully understand, retain the more accurate verbatim memories for longer.
It could, however, be some time before it’s accepted in the court room that a child’s memory can be more robust than an adult’s. “It will probably take a false conviction and then an appeal,” says Brainerd.
There’s no doubt that our memory is a complicated beast. Anyone given the task of taming it, like the police, faces a difficult job. They walk a tightrope where asking the wrong question could easily change someone’s recollection of events. And with a witness who is reluctant to relinquish information, the idea of free recall may go out of the window.

It’s not just court cases where memories are important, either. Insurance claims or even requests for asylum rely largely on memories of past events. As Conway says: “Our memory is not a record of reality, it’s a record of our experience of reality. When you understand that you begin to recognise how big the problems are.”

Andy Ridgway is news and features editor of Focus

Ask the expert

Professor Tim Valentine of Goldsmiths, University of London, on identity parades


How good are we at remembering faces?

With once-seen faces, not particularly good. But we tend to think we are because we’re good at recognising our friends.

Why is that?

It seems to be a perceptual problem as much as a memory one. I did some work on trying to match people using CCTV footage. If we have a picture of someone and a CCTV image, we are quite error-prone with faces we don’t know well.

How error-prone are we when it comes to ID parades?

One piece of research looked at about 600 ID parade line-ups. About 40 per cent of witnesses identified the police’s suspect. In another 40 per cent of cases, no identification was made. And in 20 per cent of cases they pointed to a volunteer. So we know that one in five witnesses make a mistake. It means that in the 40 per cent where the suspect was identified, there will be some cases where the witness’s memory was not reliable.

Why are video line-ups generally used in the UK rather than live line-ups?

The police felt that suspects could frustrate the process by just not turning up to a line-up. It’s also harder to get people who resemble the suspect for a live parade. [Using the video system, the police can select from a database of ‘foils’ who look similar to the suspect.]

Could we improve the current system?

What we do is pretty good, and better than other places in the world. One thing we could do is to make the person who runs the line-up blind to the identity of the suspect. They can inadvertently give clues as to who the suspect is.


Follow Science Focus on TwitterFacebook, Instagram and Flipboard


Andy is a Senior Lecturer in Science Communication at the University of the West of England in Bristol, Programme Leader of the MSc in Science Communication and an award-winning journalist


Sponsored content