This episode of Perspectives is the second in a two-part series exploring the psychology of police interviewing.
Associate Professor Doctor Kai Li Chung is Head of Psychology at the University of Reading, Malaysia. A leading researcher in forensic psychology, her research explores how psychology can be applied to improve criminal justice systems, with a focus on investigative interviewing practices.
This episode was recorded at a seminar in Kuala Lumpur on the Méndez Principles on Effective Interviewing. We resume with Dr Kai Li describing how an investigator’s presumptions about a person’s guilt or innocence can influence the conduct of an interview.
Hello and welcome to Perspectives, the APT’s podcast which explores contemporary issues related to torture prevention and dignity in detention.
I’m Almudena Garcia, APT’s Digital Communication Adviser, and we are delighted to share with you the second episode in a two-part series exploring the psychology of police interviewing.
Associate Professor Doctor Kai Li Chung is Head of Psychology at the University of Reading, Malaysia.
A leading researcher in forensic psychology, her research explores how psychology can be applied to improve criminal justice systems, with a focus on investigative interviewing practices.
This episode was recorded at a seminar in Kuala Lumpur on the Méndez Principles on Effective Interviewing.
We resume with Dr Kai Li describing how an investigator’s presumptions about a person’s guilt or innocence can influence the conduct of an interview.
Dr Kai Li Chung
So going back to this idea that interrogation is a guilt-presumptive process. So whenever you have some expectations, it will activate a behavioral confirmation process. So if you think somebody is innocent, it will influence the way in which you conduct the interview, or the investigation. As opposed to, if you think a person is guilty, it will also influence your behaviour towards that or how you perceive that case. And this is what we mean by behavioural confirmation.
So a very interesting experiment. There was again a typical mock crime where you've got people who were asked to do something, pretend to do something bad, and some who were innocent, and then they were being questioned by the interrogators and the process was being recorded, the interview was being recorded. And now, interestingly, is that they manipulated the interrogator's expectation. So they told the so-called detectives, who are also participants, they say, "Okay, the people that you will interview later on, one out of five of them are innocent”. And then they would say that the people that you're interviewing later on, four out five of them have actually committed the crime. So that they have the probability of people that they're interviewing, four out of five will be guilty. So this is the guilty expectation.
Do you think that that's going to influence how you conduct the interview or your likely outcome of what you think the interview is? If you think four out of five are guilty, what would you do? Again, we think that the police interrogation is a very simple job. It's actually a very difficult job, right? We don't know, what would you actually do? You will try a little bit harder to make sure people confess so that it matches your expectations that four out of five people are guilty. So that is indeed what they found.
So this idea, that if you believe that somebody is guilty, in the guilty expected condition, you will try harder to get them to confess the crime because your expectation is that they're guilty. In general, it's a coercive technique, a psychologically manipulative technique. And this is a very famous quote. "These methods produce false confessions because they convince innocent suspects that their situations are hopeless, just as surely as they convince the guilty that they are caught."
Again, reflecting back to the poll that we've done, you will never confess to a crime that you did not commit. Everybody thinks that it is illogical to do so. You must be stupid to confess to something that you didn't do. It doesn't make sense. We think that it's illogical, but it's actually a very, very rational decision making process because if the police officers told you that if you do not confess, you'll be deprived. I will keep you here indefinitely, you won't be able to go home, your family will be harmed, things like that. Just on the interview we watched they said “my mother told me that if I didn't do this, I would die here”. So what's worse? Go to prison or die in custody? So I might as well go to prison than to die in custody. At least if I was in prison, I can appeal to get somebody to save me. So when you are placed in that situation, confessing is actually a very, very rational decision making process. It's not that people are just illogical, it's actually decision making that is understandable. So this helps us explain why people do not remain silent. Because you want to get out of that situation. You want to ensure that you've got every opportunity to prove that you're innocent.
Okay, I'm going to just talk a little bit about this idea of compliance as well as obedience. And I think this is something very interesting within our culture, this cultural hierarchy, culture of obedience. That adults are always right, authority is always right. But if you comply with that, you are more likely to do things that are probably not necessarily ethical in your own terms. So this is what we call coerced compliance. Very psychologically oriented. And that is the reason why some people are more likely to be vulnerable to false confessions.
Okay, as you may be aware, the Méndez Principles and the PEACE model, are very much based on rapport building. And we always have this sort of question asked, “Okay, when you build rapport with people, are you trying to be sympathetic, empathic towards them?” And imagine, sometimes, and I'm sure police officers will know this too well, is when you get somebody who confessed to you about something that's really, really unpleasant. And this rapport building is a very, very difficult job. And to assume that anybody can do it is a very, very dangerous thing to do. But the idea is that when you have a lot this rapport-based training and, interestingly just for your information, is that the way in which we understand rapport, at least Malaysians, is slightly different from the way in which many Western or Europeans understand rapport building. So we have done some research in that and we are thinking that if you understand rapport in a very, very different way, it will have implications on how you conduct the rapport building part of the investigative interview.
So we have done some preliminary work looking at rapport building where, in the very, very professional sense, when you build rapport with people, you're very professional, and in a very individualistic culture, it's okay to assume some space. In our very collectivist culture, we look at things as very harmony-based. We look at things as trying to not create any conflict. So if you understand rapport in that way, sometimes if a police becomes very, very nice to you, you might think, “Do you have anything behind? Why are you nice to me?”. And we notice that in children, actually, in my work with children, when the interviewers are very nice to children, their memories are actually worse because they start to think that it's a game. But if you are very authoritative and you say, "Okay, this is a task that I want you to do, I need to remember as much as you can", they actually perform better. So there's something, again, it's preliminary data, we think that something is going on in terms of how we understand hierarchy. So my point is sometimes it's very easy for us to think that, “oh, something works well, let's just go ahead with it”. But there are some very, very interesting cultural differences that we need to take into account. This doesn't mean that we cannot do it, it just means that we need to take a little bit more of a different approach to these things.
So going back to this, so when you have these sort of soft sell techniques, minimisation, when you give your suspects a very false sense of security, you say, "oh, you know, no woman should be on the street looking like that sexy. So any normal man will do what you did." Right. So the idea is that you make the suspect identify with your views and to get them to speak. Or you scare them with, you scare them with the fact that if you don't confess, I'm going to ensure you get death penalty or something. If you threaten people with serious consequences, that is what we call maximisation. But when you imply leniency, it is also very interesting. When a police officer implies leniency, even though they have no control over what the sentencing decision is supposed to be, it is also likely to get people to falsely confess.
So one of the techniques that they recommend police officers to do is that you should present an alternative question. So you could ask them questions like, "When you first grabbed her by the side of the road, did you intend to rape her or did you just plan to rob her?" Which is a better option? Remember that after many, many of the same questions again and again and again and again in the six hour interrogation, sometimes people just settle for the less serious one. Because rape is seen to be more morally unacceptable. Or it's another thing, questions like, "Oh, is this the first time something has happened? Or have you done these things a hundred of times before?" Again, a hundred of times looks a hundred times worse. So you might just pick the one. Now I'm not saying that everybody will give into this, but it has been shown to be highly dangerous if we're using it with people with intellectual disability. If you ask people with mental disabilities, they're more likely to give in to this question. And that is why the Méndez Principles advocates open questions. It applies to suspect interviewing, witness interviewing, victim interviewing. Which is true, because if you ask a victim of a sexual abuse – for instance, “did your perpetrator wear a green shirt or a blue shirt?” – what they have found is that somebody with an intellectual disability, they are more likely to simply pick an option. Not because they don't know but they think that you want them to pick an option. So that's why we advocate open questions. So instead of asking if they wear blue or shirt red, “what color shirt did the person wear?”. And there's sort of guidelines in terms of how you can do the questioning better, because they reduce the risk of unreliable information.
Okay, so what if you ask the public what do they think about false confessions? So this was a study that was done by Henkel and colleagues in 2008 and they wanted to understand what are people's attitudes and beliefs towards false confession. This is what they found. Well, before what they found, I wanted to mention this quote by Kassin. "Reasonably, most people believe that they will never confess to a crime they did not commit, and they cannot imagine the circumstances under which anyone else would."
So it's very much similar to what we discussed at the start. Nobody would confess to a crime that you didn't do because that's just illogical. But if you ask people, and you ask big numbers of samples, a big sample size, a confession is a strong indicator of a person's guilt. Majority of people agree, right? That if you confess, you must be guilty. So a majority of people believe that if someone has confessed to a crime, they're probably guilty. About half of the sample believe that. Only a small subset of people – people who are mentally ill, people with psychological problems – are vulnerable to false confession. A very huge number of participants believe that. So this is all things that we acknowledge. Interestingly, this means that people acknowledge that false confessions can happen. But what is interesting is when you ask them, “do you think you would?” Again, similar to the findings what we had just now, overwhelmingly people say it can happen to other people, but it won't happen to me.
And this is what we call a fundamental attribution error, where when some things happen to somebody else, you would think that it happened to them because of circumstances within them. Maybe they're not very smart, maybe they were not very wise, they were not very rational. But it won't happen to me because I'm smart and I'm rational. So we have a tendency to attribute somebody's fault to something about them. But if it happens to you, what you would say is that, "Oh, it was circumstances that caused me to do it. It was not because of me. They really, really made me convinced that I was that". So there's this idea that if something bad happens to somebody, it was because they asked for it. But if something bad happens to me, something caused it, not because of me.
The study also looked at people's perception of the likelihood that a suspect is guilty based on the available evidence. So again, you can see that if there is evidence, DNA evidence, about 80% would think that the person is guilty because it's hard evidence. If a person signs a confession, 60% people think that you're probably guilty. Again, this idea that if you sign a confession, there's a high chance that you must be guilty. But we know through case studies and through the examples that I've shown you just now, that sometimes it can occur even though it does not indicate their guilt. And, so most people, again, continue to believe that an innocent person will not falsely confess unless they're physically tortured or they're mentally ill. So what can we do?
Electric recording of an interrogation is a situation where everybody wins. I know in Singapore, I've worked with someone with the Ministry of Home Affairs, where they have implemented compulsory recording of suspect interviewing. In Malaysia, the Child Interview Center recordings do happen with the children, but I don't think it's mandatory for suspect interviewing. In Singapore they had some qualitative work where they interviewed police officers after their mandated interviewing and they actually found quite positive feedback. Even police officers are quite pro-recording because they said, “well, at least now I don't have to defend that I didn't use any sort of dodgy techniques or so on”. And also, interestingly, we did some work with a very, very small sample size of police. We only had 40 participants, but we noticed about, I think 60%, the majority of people, were okay with recording interviews. But again, this is a very bottom-up approach where we ask investigative officers, while people who are more in the position of authorising these things may not share the same ideas. But it's interesting that it actually, recording is fairly economical nowadays, but it's not implemented all over the world yet.
Okay, for now I'm just going to move on to a little bit. I've talked about all of the dodgy things that can be done. So moving forward, what can we do? You might have heard about the PEACE model. This is based on some psychological research. I'm not going to talk about this in a lot of detail, but it involves planning and preparation, engage and explain, account, closure, as well as evaluation. So then I'm just going to the data, is it effective? Right? That's what I want to know. Do the Méndez Principles work? And the Méndez Principles are based loosely upon the PEACE model and the work that has been done on the PEACE model.
So Dave Walsh and Ray Bull, so this is from England, they examined about 142 interviews of people who have committed crime and they wanted to know whether or not there's any relationship between how skilled you are in interviewing and outcomes of the interview. So again, we talk about how the PEACE model is based on planning, engage, account. There's a closure, there's evaluation and it's a lot. So rapport building that goes on within the engage and explain phase, and account is where you want the suspect to give you their account. Their account, not you assuming what happened, but them telling you their version of the story. And what they found was that if people were better at administering the PEACE model interviewing, it was actually associated with a higher number of comprehensive accounts. And it's also associated with more admissions and confessions. So it's something that we want, isn't it? As people working in the field as police officers, this is a KPI we want to achieve. Better outcomes if you did interviews in a better way.
Some studies, they asked, they wanted to know, what are the strategies used and how do the suspects respond? And these are all analyses of real life interviews, taped interviews with suspected murderers and rapists. And interestingly, there were certain types of components of the interview that made people more likely to confess. So if there was rapport based, if it was based on empathy, open type questions, it was associated with an increased likelihood of admitting to the crime. But if they had a lot of negative questions, implying that they were unpleasant people, something really, really negative, it was associated with decreased likelihood of admissions.
Same thing found in Sweden, where they were asking murderers and serial sexual offenders about their experiences and how did the offenders respond to their interrogations and, in general, the people being interviewed perceived police officers to have two main types of approaches: a very dominant approach or a very humanitarian type approach. And when you account for these things using statistical methods, the dominant type of approach is associated with more denials. So again, showing that being coercive, being dominant, being unpleasant is not leading to better outcomes. If anything, it is leading to more negative outcomes. And the humanity approach is associated with more admissions.
All this is European, Western data. So what do we know about the data from the east? We don't know a lot about this regional work here. Again, it's a work in progress. But in Japan they did a similar study and Walsh and colleagues, they asked police officers about their experiences interrogating suspects, what other techniques that they use. And they noticed that, again, when the interviewers employ a relationship-focused style, the suspect was more likely to make full confessions. They also looked at the personality of the officers. Who are the people who are more than likely to get better outcomes? And the people who were engaged in active listening, rapport building, discussion of the crimes, they were more likely to obtain full confessions and also more likely to gain more information from the suspects. As opposed to using very confrontational techniques. And the more confrontational they were, the more likely the suspect would just give very, very vague answers and avoid eye contact. So it seems to suggest that an appropriate use of empathy, not asking you to be friends with the suspect, it is appropriate use of empathy, trying to understand it from their point of view. It's recommended in the Méndez Principles and it's found to be effective to get people to provide investigation-relevant information. Again, there's some work that looks into whether or not there are certain type of police officers that are better at conducting certain type of interviews. And again, the idea of whether or not people can be trained to conduct better interviews. It's sort of a very messy literature. But I think the basic thing that authorities can provide is to provide some appropriate training so that they can do a better job.
So these are all interviews from the police, but very, very controversial. Not controversial. I mean, an interesting study that is done in the USA, and I say controversial because it's very difficult to get approval to do this from an academic perspective. You get prisoners to answer questionnaires about the techniques that police officers use. So in America, they did this study and they found that when they asked prisoners what do you think an interview, police interrogation should look like? And did you intend to deny or admit to your crime? And this is a very interesting stance. So they found that 39%, they already made up their mind before they went into the interview that they were going to deny it. So these are the people where the techniques are supposed to gain more information. So the Méndez Principles apply. But look at this, interesting, 36% entered the interview not yet deciding whether or not they want to confess during the interview, the fence sitters. So again, what is interesting to know is how do you get them, if they were guilty, again, we cannot presume that they're guilty, but if they were guilty, how do you get them to confess? And if they're not, to not make them confess because that would be a false confession. But this also means that 25% of the jail inmates have already decided to confess anyway. So what's the point of going through the whole coercive technique if they were planning to confess anyway? So again, having proper planning of the interview, having a more sort of ethical method of interviews may actually serve a better purpose. And so these findings show us that the assumption that a lot of us have, that people will always deny accusation, is not necessarily true. Again, it is a misconception.
And some suspects have said that they were waiting to see how they were being treated before they decided whether or not they wanted to confess. So again, humanistic approaches may lead to better outcomes. But in general, it's difficult for people to accept that they could do better. And we have to acknowledge that. So how do you approach these methods without being upfront? You're rubbish, it could be better. And that's a technique, isn't it? We have to be better with that, persuading people. And I think that's probably useful as we try to disseminate the information here. There's also practical implications, which I understand, that sometimes a lot of authorities, their hands are tied. There's financial resources, you need expertise. Because to do quality training, it requires resources, time. You're asking police officers to take time out of their already day-to-day, very, very tedious jobs to do a training. It's not easy. And I talked about biases and misconceptions, this incorrect common belief, this idea that, that torture works. “I know it because we've done it before and we've got confessions”. But again, the misconception that people who are arrested and in the interview room are definitely guilty. And we talked about the biases that we may have. And also, I think the lack of this empirical data locally to support how effective this is, and I touched upon this just now. It's all nice and good that we are wanting to implement the Méndez Principles, but you have to convince our local community that it will work.
So just a very, very final point. To summarise what I've talked about, people can be induced to falsely confess. And there are a range of factors that can motivate this to happen. But my take home message is that a respectful, open approach is actually a win-win situation. And we need to be open-minded to adopting some of these principles. And it takes a lot of work from all parties. And I'm so grateful that all of you are here today. So that's all for me. Thank you very much.
Associate Professor Doctor Kai Li Chung is Head of Psychology at the University of Reading, Malaysia.
The transcripts for both episodes in this series on the psychology of investigative interviewing are available for download.
Thanks for listening to Perspectives and we look forward to your company next time.