Lie Detection:
What is best – A jury or scientific examination?
Trial by Jury is for most people a basic civil liberty and protection against unfair conviction. This article examines recent developments in forensic scientific lie detection and provides information in support of the writer's contention that forensic lie detection can and should provide vital information of relevance to issues affecting prisoners.
This article looks at forensic developments involving FMRI (Functional Magnetic Resonance Imaging) scans as well as polygraphs and a new form of lie detection known The TARA (Timed Antagonistic Response Alethiometer). The TARA is a new type of lie detector which involves a computer test designed to calculate a reaction time for cognitive processing. The theory is that liars have a harder task to undertake than truth tellers and that this time lapse is detectable.
Over the years polygraphs have been criticised for various reasons, including the fact that they are awkward and expensive. It has also been suggested that polygraphs could detect whether or not someone was telling the truth, or in other words to discover whether or not they were liars. It could be argued that it would have been better if those advocating the use of the test had accepted the fact that absolute proof of the telling of lies could not be provided by a polygraph test. Arguably it would have been better had advocates of the test limited their arguments in favour of the test to discovering whether it was likely or not that lies were being told.
An issue in relation to polygraphs arises if by being able to stay calm a person could beat the polygraph test. The opposite of this is if it were possible for test results to be worthless if blameless individuals became stressed on the basis that he or she thought that they might in some way be implicated in crime. Efforts have been made to try to overcome these issues and to do so researchers have tried to avoid asking questions about guilt. They have sought to avoid asking simple questions such as, "Did you do it?" or "Did you kill X?". Researchers have sought to be more subtle in their questioning. They have developed what is referred to as Lykken's Guilty Knowledge Test. This involves asking questions where in theory specific knowledge would only be available to the criminal and those investigating the case. Instead of simply asking the person whether they stole an item, they could ask them to comment on where the item was. The results could then be compared. An alternative line of enquiry might, for example in a murder case, ask whether the body was found on the floor, on a chair or on a bed. The theory is that the criminal would know and would respond differently in answering the questions. Someone who had nothing to do with the crime could reasonably be assumed to respond the same way to all the questions. The whole purpose would be to seek to identify the criminal by seeing their reaction to questions put to them.
It is claimed that the Guilty Knowledge Test detects concealed recognition of crime details 80-90% of the time. What is also encouraging is that it is claimed that it incriminates the innocent in only up to 10% of cases.
If no-one is suggesting that the test is foolproof, my view is that evidence of this kind is better than leaving crucial questions to ordinary people, like you and me, who have been shown to have no particular ability to detect the telling of truth or lies. We are all clearly fallible and my view is that the time has come to use up to date technology. A test for lie detection need not necessarily be seen as deciding once and for all whether lies were being told. The point is that scientific advances can provide information in relation to the likelihood of the truth or lies being told. I believe that the difference is highly significant and that serious consideration should be given to the use of any scientific information which can assist a jury, Court or panel, such as the Parole Board, in deciding any issue which involves an assessment of truth telling when that issue is of central importance to the case.
A case in question could be in relation to the assessment of the risk which a particular prisoner poses in relation to reoffending. What a prisoner may say in relation to his future plans or reaction to stimulation, for example pornography, could be assessed and, in my view, the assessment could benefit from scientific input.
A crucial issue arises in relation to whether or not jurors or others are capable of determining whether or not lies are being told. The problem is that lying is commonplace and scientific evidence suggests that people's intuition in relation to lie detection is very poor. Research has suggested that intuitive lie detection is no better than 54% overall. There appear to be no groups of people who are identifiably better at deciding whether or not lies are being told. In my experience, police officers and prison officers believe that their experience allows them to be a better judge of whether someone is telling the truth. Research suggests otherwise. If we are not very good at telling whether or not someone is telling the truth, then surely there is an argument for looking to see whether scientific evidence can help get to the truth. Research suggests that we can detect the truth in about 65% of cases and in about 45% of cases in relation to the telling of lies.
Polygraph testing does not appear to be the answer to the simple question of whether someone is lying or not. However, as previously suggested polygraphy could provide information in relation to the likelihood that someone is telling the truth or telling lies. The trouble with polygraphy is that it is put forward as a means of telling truth from lies, and cannot guarantee that distinction. Perhaps as a result of the way it has been marketed, polygraphy has largely been discredited. A serious criticism is in relation to how it can create excessive false positives and how it might be used to elicit confessions.
If polygraphy is not the way forward, there is also the possible use of FMRI testing (brain scanning) where the Computer tests reaction times, the "TARA". Various lawyers at Swain & Co have had the opportunity of speaking to Dr Aiden Gregg who is involved in the development of the TARA lie detector. Dr Gregg is an experimental psychologist at the University of Southampton. We have discussed with him the operation of the TARA and differences between the TARA and other lie detection methods. The TARA is a new type of lie detector and the theory is that when answering questions, liars go slower than truth tellers. The theory is that it is more difficult to tell lies and that telling lies involves a different part of the brain. This is what the FMRI scans seek to detect. We discuss later in this article how these work. The TARA has been shown to provide reasonable accuracy in relation to lie detection and a figure of 85% is being quoted. It is accepted that stress may affect the results in any form of lie detection but in the case of TARA it would appear that if the person being tested is under stress, this may increase accuracy rather than have the opposite effect.
TARA does not set out to state whether or not someone is telling the truth. It is not held out as a means of providing a yes or no answer. It deals with degrees of likelihood or probability. This is of course a well known concept for lawyers. In criminal cases the burden of proof is beyond a reasonable doubt. In the case of civil law the test is on balance of probabilities; more likely than not. Findings of fact by the Parole Board at Oral Hearings are also made on the balance of probabilities. The Courts are, therefore, used to dealing with less than certainty and dealing with degrees of probability. One advantage of TARA is that this is exactly what it sets out to deal with. It provides information in relation to the relative probability of whether the truth is being told and at the same time the relative probability of whether lies are being told. The faster the response time, the more likely it is that the truth is being told.
Dr Gregg argues that TARA could be used not only in Court but could be used for other purposes such as deciding where to focus criminal investigations. TARA is being considered for use by the Home Office to authenticate claims of sex offenders on probation. I believe that this investigation is appropriate on the basis that it may be a cheap and effective means of supporting truthful information supplied by prisoners. I see no reason why consideration should not be given to using TARA or other scientific lie detection systems with a view to assisting in establishing hotly contested issues involving establishing what is the truth.
Finally, there is the possibility of use of FMRI scans to assist in such issues. This is currently being pioneered by various experts including Professor Spence of the University of Sheffield. FMRI is at the cutting edge of brain scans and investigations have discovered that lying activates brain regions involved in suppressing information and resolving conflicts. The point is that telling lies involves a different part of the brain. The theory is that appropriate use of an FMRI scan can help get at the truth despite a persons attempts to hide it.
Functional Magnetic Resonance Imaging (FMRI) involves scanners measuring localised brain activity by determining blood flow and oxygen utilisation in portions of the brain. The aim is to identify activity associated with a particular area of the brain. Research suggests that lying appears to involve brain areas associated with high level executive functions, which are required to suppress what would otherwise be a true reply. Issues remain in relation to the risk of obtaining false positives.
There is likely to be considerable judicial resistance to the use of any forensic lie detection procedures. Matters are not helped by evidence of some countries having had bad experiences with polygraphs. A crucial issue arises in relation to whether the tests be considered to amount to scientific evidence and as such be admissible. There are issues relating to reliability and validity.
Assuming that test results were considered to be scientific evidence and admissible, I anticipate that witnesses seeking to bolster their credibility would want to produce such evidence. They may seek to use this evidence for screening purposes and as part of any interrogation process this would raise serious human rights issues. In the USA this is covered by the Fourth Amendment's prohibition against unreasonable searches and seizures.
I guess the conventional argument at the moment would be that the judiciary would be wise to resist the temptation to admit scientific data into evidence before more is known in relation to reliability and validity. There would also be safety issues. However, it is difficult to perhaps resist the use of lie detection when it is not directed at providing information in relation to guilt. It could solely be in relation to risk e.g. in connection with a parole. There could be a value in relation to the most difficult cases. The agreement of participants in testing would be essential and advice would need to be given in relation to the medical, ethical and legal issues arising. Providing this advice and drafting written information would be highly problematic but in my view would not preclude efforts to find these matters.
People in favour of FMRI scans have argued that they could be used in relation to solving crimes, in identifying military spies and perhaps in dealing with medical problems which involve compulsive lying, for example in anti social personality disorder.
It should be accepted that we all tell lies at some time. The telling of lies can make it easier for people to get along in society. For example, people are polite despite differences with a person they work with or associate with. Over the years most people have thought about whether lies can be detected through body language. Polygraphs deal with tracking a body's responses on the basis that if someone is not telling the truth you will typically see that the person breaks out in a sweat and shows a brief heart rate drop. Others have thought that nervousness or for example touching the nose may be indicative of not telling the truth. The problem is that research suggests that we all have difficulty detecting both someone who is telling the truth and whether someone is telling lies. It seems clear that there are no telltale signs that reliably show whether someone is a liar. Indications of arousal can perhaps do no more than give information in relation to the possibility or probability that the truth is not being told. Use of FMRI scans suggest that lying activates brain regions involved in suppressing information and dissolving conflicts. In appropriate cases it could perhaps be argued that there would be benefit in undertaking polygraph type tests combined with brain imaging. I would argue that each could provide valuable information in relation to the likelihood that someone is telling the truth. The TARA scheme provides further information and also has the advantages of being relatively inexpensive and non invasive.
I consider it to be about time that forensic lie detection was trialled or actively pursued in relation to the assessment of risk in relation to parole as the Parole Board makes findings of fact on the balance of probabilities. The criminal Courts make judgements of facts when they are satisfied beyond reasonable doubt. For this reason, the Courts are not perhaps yet ready for the use of lie detection in relation to guilt. However, I believe it to be only a question of time before evidence of this kind is put before the Criminal Courts. If intuitive lie detection is very poor, I believe that this is something which Courts should be commonly reminded of and that in the 21st Century it make sense for technology to be used to provide information not in relation to guilt or as a categoric detector of truth or lies. I believe that the time has come for scientific evidence to be used in relation to the likelihood of whether the truth or lies are being told. I believe that this sits comfortably with the burden of proof in criminal and civil proceedings. The burden of proof in criminal cases is to prove a case beyond a reasonable doubt. In a civil case the burden is to prove a case on the balance of probabilities; more likely than not. I believe scientific developments can provide vital information in relation to the likelihood that something is true or false.
As a lawyer I believe that one of the greatest challenges or opportunities available to a lawyer is the opportunity to champion the use of technology in the Courts, in particular in cases where crucial issues arise in relation to “his word against hers” or similar. The ultimate decision would not be taken away from the jury or from the Judge. Science would simply be providing information in relation to the likelihood that either truth or lies were being told. This information would be presented to the Court as another piece of evidence for it to consider when deciding a case. In this regard it would be no different to other forms of scientific evidence used by the Courts, such as DNA or fingerprint evidence.
The writer is grateful to Dr Aiden Gregg for information supplied in relation to the TARA lie detector and Professor Spence for providing information in relation to the use of FMRI scans. Readers may wish to also refer to Scientific MIND volume 20, number 1, page 50 which deals with the use of FMRI scans and polygraphy.
Graeme Swain is a partner in the firm of Swain & Co solicitors, which specialises in prison law and criminal law. The firm has offices in Southampton, Havant, London and Liverpool and provides a national service to prisoners.