‘Psychology has been fundamental in terms of informing witness interview practice’

In the first episode of the Psychology Unleashed podcast from the University of Westminster, Professor Catherine Loveday hears from Professor Coral Dando, Professor of Forensic Psychology at the University of Westminster, about eyewitness memory and more. Listen in full here, or enjoy this edited transcript.

You started life as a police officer and not an academic, didn't you? What inspired you to become a forensic psychologist? 

I was a police officer working in central London for just over 10 years. That brought me into contact with the general public, particularly when they're involved in criminal justice processes. And in particular, it was about understanding eyewitness memory and how as an investigator, I could help eyewitnesses to remember more. It was really apparent during the process of interviewing eyewitnesses and asking them questions that they weren't necessarily remembering as well as I'd expected that they could, or that they should even… I couldn't understand why someone that had seen a violent and unpleasant incident couldn't tell me in a lot of detail what had happened and who had done this to them.

So you perhaps had some of those common preconceptions around memory – that people should just be able to report stuff that's happened. Your work has taken you in a different direction… how have you changed what you believe about that?

You're absolutely right. As a police officer in my late teens, early 20s, I didn't know anything about how human memory worked, and I was of the opinion – and probably in many areas, that opinion still prevails, unless you are told otherwise – that human memory is like a tape recorder, that we absorb absolutely everything and at a later stage, when we're asked about it, we can just churn it all back out again, and all that information will be absolutely spot on. Of course, what I now know is that doesn't happen. 

What I started to do as part of my PhD was to understand how we can help witnesses, victims, survivors of crime, to remember as much information about what they've experienced and to make sure as far as we can that the quality of that information is good, so that they don't make too many errors and that they don't feel duty-bound to give you an answer even when they're not sure. We know that the way in which we ask questions, and the way in which we manage the social interaction that occurs between an interviewer and an eyewitness, can support eyewitnesses to remember more about their experiences. 

It's so good that Psychology is providing the evidence to actually change some of these practices, and I know that's something you've been working on. Can you say a bit more about the concept of 'rapport'?

An interview is a social interaction. Interviewer and interviewee have to engage with each other on a social level, and that has to be comfortable for the interviewee. We don't have an infinite amount of mental energy. Imagine having to split some of that into trying to manage the social interactions – I'm looking at what I say to you and how you receive that information, I want to do really well, I want to remember as much as I can. My memory might not be very good. I might be a little bit embarrassed because I can't remember details about the person that I saw. I'm monitoring how you're receiving my responses, and if that's an uncomfortable social experience – if the interviewer is looking a bit disappointed or raising their eyebrows because I can't remember as much as they would like – then that detracts from my ability to actually spend time thinking. 

So we know that rapport building is really fundamental to maximising eyewitness memory. We've got this little mantra – a comfortable eyewitness tends to be a 'good' eyewitness. They can spend more time thinking and remembering the event; they feel more comfortable explaining when they don't know or they can't remember information fully; and as a result, this whole interaction feels more comfortable for the eyewitness. That's what rapport building does. 

It's so instinctive, isn't it, to believe that the fuller and richer the account, somehow the more accurate it is. But what you're saying is that what's most important is that a person feels able to say, 'Actually, I don't remember', and to acknowledge the gaps in their memory.

Absolutely. Saying 'I can't remember', or 'I don't know', or 'I didn't see', is an indicator of a comfortable eyewitness, because that person is comfortable. It's saying to me, I don't know, I can't remember, I didn't see.

We've talked a bit about how you carry out the interviews, but is the timing also important? 

We know from the psychological literature that the quicker we can get someone to 'practice' what they've seen, to provide an initial account, the more robust the memory. It becomes stronger and more resistant to interference. And we know that over extended periods, that can strengthen the memory trace. But of course, in a criminal justice setting, it's not always appropriate to collect information from a witness straight away. And in any case, a witness interview is quite a long and in-depth process. It's not a quick 'question/answer' format. 

So how do we get around this? 

That's a good question, because we have just completed some research on collecting what we call an initial account… in non-professional environments, I refer to it as a 'quick and dirty' account. It's using what we know about how to support individuals to give us good information by selecting types of questions we ask. We make sure that they are not leading, they're not forced choice, and they're asked in an appropriate manner. But we do that quickly. That might be at the scene of a criminal event, if that's appropriate, or it might be within half an hour, or it might be within an hour. An investigative witness interview is not going to happen in that time.

There's a way of doing that, using in-person, face-to-face questioning, normally by a police officer, often on the side of the road or at someone's house. It's quick, straightforward, operationalised: the questions are very fixed format. We are working here at Westminster at taking that format and putting it into a chatbot which individuals can access where it's appropriate – it won't suit everybody – and provide their answers immediately and textually, rather than verbally. Because it's such a specified initial interview protocol, it ports itself across to chatbots quite easily.

That's great. And presumably that's also useful if you have a large number of witnesses, where you can't possibly get around them all at once. 

Yes. Where there's a large event, you could, in theory, send out a link to the chat bot, which we've referred to as 'Chat Charlie'. It's been developed to ensure that your responses are matched by relevant replies from the chat bot itself. It's conversational in nature, but not like the bots we come across through banks and customer service – it's a very fixed process. It can be sent out quickly, the information collected is stored securely and is available quite quickly to investigators, in a way that it might not be if you did an in-person initial account. There might not be enough manpower to do that.

That could make such a huge difference… if you had, say, 100 witnesses, even if it was only suitable for 50 of those witnesses, that's 50 accounts who've gone through this initial, quick and dirty account, which means that theoretically they should then be able to understand better when you come to do a full interview. 

That's right. An initial account isn't designed to replace a full investigative interview. It's about underpinning and strengthening the work of investigators. But psychologists have also recognised it's got cognitive benefits, because it introduces that practice effect. It might even reduce disinformation or interference from things such as Instagram or media that often post footage. We know eyewitnesses can be manipulated by external images… so that's possible, but it's moving into the future a bit.

But in this world where we are so quickly exposed to different people's images and accounts of what's happened, it's so important to have captured that first account. Chat Charlie sounds like a good way to do that. But as you said, it's not appropriate for everybody. We know from memory research that some people find it harder to give an account, and I know that you've been working hard on trying to address that. 

Yes. If we move away from an initial account now, to talk about the witness interview, which often happens at a police station or at a person's home or other appropriate interviewing facility, the way in which we collect information during that more formal process needs to alter according to the type of witness that we have in front of us. Psychology has been fundamental in terms of informing practice in that regard. 

There are obvious differences between types of witnesses, such as age. We know that children's cognition is very different to adults, and that their long-term memory is not fully developed, and also their understanding of questions and how to explain what they've experienced is different. But some differences aren't as obvious. A classic example of a group of individuals I've done a lot of research with is those with autism. They are living in the community quite happily managing their autism diagnosis, but people with autism are often over-represented in the criminal justice system, both as suspected offenders, and also as witnesses and victims and survivors. When they come to us wanting to tell what they've experienced, we have to interview in a different way. One way is to help them to do a type of mental time travel, to think back to the event and to draw whatever they remember about it.

That sounds like my worst nightmare. I'd hate to have to draw something… that sounds quite difficult!

You're right. And a lot of people will say, 'I can't draw it'. And of course, if people aren't comfortable drawing, then they don't. But my response to that is, 'I don't want you to draw what happened. I just want you to draw what reminds you about what happened'.

So you're not expecting sort of a Van Gogh production of what they saw. You're just trying to get them to express themselves in a different way.

Yes. I explain this is about triggering the cognitive processes that underpin searching through our memory – going through the filing cabinet, looking for all the elements that we want to bring together to remember that event. And generally, the drawings are quite abstract. There can be shapes, stick people, bird's eye views of places. They are all fundamental in triggering the memory process.

It's such an intriguing way to do it, and it's interesting that it works so effectively as well. And you're training people to start using this in the real world now.

Yes, we call it 'Drawing to remember'. There was a recent Ministry of Justice guidance document – Achieving Best Evidence in Criminal Proceedings, a big, thick guidance document on how interviewers and investigators should go about supporting vulnerable individuals. Drawing to remember is in there, which is really exciting… it's good evidence of research to practice. Police officers are trained in that technique in the UK, but also now in many countries in Europe who are following our lead. We are leaders in the area of interviewing. 

I've also done some training with speech and language experts in the NHS who work with children, adolescents and young adults who are finding it difficult to explain what's happened to them. The drawing often supports them, not only with their memory, but also the social interaction. If we go back to thinking about the fact that an interview is a social interaction – a conversation with a purpose – drawing helps individuals with autism to focus away from the person and concentrate on their drawing. And attention is drawn away from challenges with verbalisation and to the drawing. When you get that more relaxed atmosphere, it helps them to verbalise their experiences. 

We've also used drawing in circumstances where individuals who are neurodivergent have been suspected of being involved in criminal activity. It can help them to just describe, or not, their involvement. That research is at its early stages.

So Coral, it's been so interesting to hear about these ways in which your work is changing the criminal justice system for the better. You're changing so many things, but if you could just choose one thing right now that you could change right across the criminal justice system, what would that be?

It's got two parts to this reply… sorry, I know you said one thing! It's like Desert Island Discs… 

For me, as a cognitive psychologist that works in the criminal justice system, the gold standard would be the questionlesss interview – a way of helping people to tell what they've experienced without asking questions. As soon as you start asking questions, we introduce what we call demand characteristics, which is a posh way of saying the social interaction gets in the way. It introduces errors, a lack of confidence in the ability to remember, feelings of inadequacy if I can't tell you as much I want, and that filters down. 

The second element would be having a suite of techniques that we could pick out of a toolbox and use according to who sits in front of us. That would be my main one… we want a toolbox of appropriate techniques that people can access to improve and countenance access to Criminal Justice.

• Find more on the podcast here.