Throughout history it has often been assumed that lying can be detected by examining changes in bodily activity - but we are actually deceiving ourselves if we believe there will ever be an error-free way of detecting deception. Polygraph tests in particular should not be ascribed special status.

I make this conclusion as the chair of a working party convened by the British Psychological Society to examine research into the most popular polygraph tests and assess their use in real life situations. The working party’s report, which was published in January and was entitled “A review of the current scientific status and fields of application of Polygraphic Deception Detection”, concluded that the accuracy of polygraphs is not high and that the rate of incorrect decisions is too significant to ignore.

Of course polygraph tests are not currently used in criminal investigations in the UK but they are in many other countries including Belgium, Canada, Israel, Japan, Turkey, Singapore, South Korea, Mexico, Pakistan, the Philippines, Taiwan, Thailand, and the USA. In a number of countries the courts have been apprehensive about admitting testimony concerning the ‘outcomes’ of polygraphic lie detection and the BPS report should be of assistance in this regard.

Polygraph tests work by measuring changes in bodily activity such as heart rate, blood pressure, respiration, and palmar sweating. Three out of the four most popular lie detection procedures assume that while answering so-called ‘relevant’ questions, liars will be more aroused than while answering so-called ‘control’ questions.

Yet this premise is somewhat naive as truth tellers may also be more aroused when answering the relevant questions, particularly when these relevant questions are emotion evoking questions, for example an innocent man, questioned about murdering his beloved wife, might experience strong feelings about her. An innocent examinee can also become more aroused due to fear which may occur, for example, when the person is afraid that his or her honest answers will not be believed.

Moreover, a suspect may admit having guilty knowledge but nevertheless deny guilt. This happens when the suspect admits being present but denies the specific alleged acts, for example in an alleged sexual assault where the suspect admits the sexual acts but claims that they were consensual.

In our report we examined the available evidence to establish whether the polygraph is a useful procedure for the UK. Scientific laboratory studies, which generally show somewhat favourable results for polygraph testing, are strongly attacked by polygraph opponents. Amongst other things, they argue that the ‘guilty’ participants, who are asked to commit a mock crime, have little incentive to try to beat the polygraph test and that innocent participants are unlikely to be concerned about the relevant questions.

Field studies illustrate the accuracy of polygraphs in the ‘real world’ but their quality is subject to debate. One of the main problems is establishing with certainty whether the suspect is actually innocent or guilty. Confessions are widely accepted as ways to establish the ground truth, however a guilty suspect who passes a polygraph test is unlikely to confess and with no confession the incorrect polygraph decision will not be noted.

Most field studies have been carried out using the Control Question Test (CQT) technique, which compares responses to specific questions about the crime (relevant questions) with responses to control questions, which are expected to arouse anxiety but to a lesser extent than the relevant questions. Overall field studies show the CQT polygraph technique catches guilty suspects in 83 per cent to 89 per cent of cases. But innocent suspects do less well with between 11 per cent and 47 per cent being classified as guilty.

The two field studies conducted using another polygraph technique, the Guilty Knowledge Test seem, revealed very good results regarding the classification of innocent suspects (94 per cent and 98 per cent of innocent suspects were correctly classified) but rather poor results regarding the classification of guilty suspects. In these two tests only 76 per cent and 42 per cent of guilty suspects were ‘caught’ and correctly classified.

This is probably because some guilty suspects may be able to `cheat’ polygraphs by suppressing their physiological reactions with the help of countermeasures. Mental countermeasures include meditation, training in hypnosis to produce ‘amnesia’ for the offence, and biofeedback training. Guilty people can also use physical countermeasures such as using drugs prior to the examination to dampen physiological responses or increasing their arousal on control questions by inflicting physical or mental pain on themselves or producing muscle tension. This reduces the differentiation in bodily activity.

Proponents of the polygraph test argue that it is highly improbable that countermeasures can succeed because properly trained examiners would notice that the examinee is trying to fool them. However, several studies, some conducted by polygraph supporters, have shown that the use of countermeasures can be very effective in defeating polygraph tests, and that they sometimes remain unnoticed by polygraph examiners.

One of the most famous countermeasures test was conducted by Floyd ‘Buzz’ Fay, a man who was falsely convicted of murder in the USA on the basis of a failed polygraph examination. He took it on himself to become a polygraph expert during his two-and-half years of wrongful imprisonment. He coached 27 inmates, who all freely confessed to him that they were guilty, in how to beat the control question polygraph test. After only 20 minutes of instruction, 23 of the 27 inmates were successful in defeating the polygraph examination.

After studying such evidence as this we concluded that even in the most favourable circumstances polygraphic lie detection accuracy is not high so an over-reliance on an imperfect procedure may lead to undue relaxation concerning the developing of: other methods of identifying or screening wrongdoers; and other ways of ensuring security and preventing crime. The belief that people who ‘pass’ a polygraph test are, therefore, cleared of suspicion is a false belief.

There are also issues about using the polygraph test in clinical settings, especially as the government has recently announced, as part of the Management of Offenders and Sentencing Bill, it is extending the use of polygraph tests for sex offenders on licence.
It is well known that sex offenders typically minimise the nature and extent of their sexual offending, probably because of their feeling of shame, so it is not surprising that the polygraph has begun to be applied with this population of offenders to aid their management and treatment.

A British pilot project with sex offenders, mainly child molesters, published in 2002 examined whether the introduction of polygraph tests assists offenders in avoiding behaviours which could increase their risk of re-offending; for example, using pornography, or going to places where potential victims might be located.

116 sex offenders were approached to take part in the study, with 50 actually agreeing to participate. Three months after the agreement a CQT was administered to 32 men (due to drop out rates). 31 of these (97 per cent) admitted to having engaged in at least one high risk behaviour, for example masturbating to deviant fantasies, obtaining pictures of children for the purpose of sexual arousal or having unsupervised contact with children. The participants were then told that because of their reporting of high-risk behaviour they would be polygraphed again in six months’ time. Out of the 32 participants who were polygraphed, 21 (66 per cent) turned up for the second polygraph test.

The researchers concluded that the polygraph could contribute to treatment programmes as a ‘truth facilitator’ rather than as a lie detector. The test was able to bring worrying behaviours to the attention of supervisors and treatment providers, allowing effective intervention and additional treatment before offending occurs.

But the results of this pilot study need to be treated with caution in view of the small number of participants, especially as it is evident that there are high refusal and drop out rates. What would happen to prison inmates and other sex offenders who refuse to co-operate with the polygraph testing? Would this refusal hinder their being released on licence back into the community? It is possible that a failure to co-operate with the use of the polygraph or a negative result (i.e. apparent deception) may have negative consequences for the sex offender concerned.

Our report concluded that the validity of the application of the polygraph with sex offenders as part of their supervision, management and treatment is still not scientifically established, although there is some growing evidence that it encourages offenders to disclose their deviant thoughts and actions, and may help them exercise self-control. Therefore, it may help with relapse prevention, but more research is needed to ascertain its effectiveness in practice.

Our report demonstrates that over-confidence in the ability of any procedure designed to detect deception, especially polygraph tests, can have serious consequences. We hope those who may be in a position to consider the use of the polygraph tests will make themselves fully aware of the contents of the report.

The report can be downloaded free of charge from the BPS website at www.bps.org.uk