There’s limited evidence that lie detectors will be effective on terror offenders
“In the case of sex offenders, [lie detectors] are a very useful way of assessing the quality or otherwise of the assertions being made by offenders...”
“…the use of lie detectors has proved to be pretty effective in giving some support and either reassurance to assertions being made or to contradict some of the claims being made by people who seem to be superficially compliant.”
Robert Buckland MP, 21 January 2020
The government announced plans this week to make terror offenders spend more time in prison. As part of that, it wants to introduce lie-detector tests for convicted terrorists in the probation system. This has sparked debate over whether, in fact, these kinds of test are actually effective.
Lie detection—or polygraph—tests are already used in England and Wales for certain sex offenders on parole, and they’re being trialled for use on domestic abuse perpetrators. The application of these tests to sex offenders were first trialled as early as 2003, and the law was changed to allow mandatory testing in 2007 and 2014.
There isn’t enough evidence to prove the Justice Secretary’s claims about the effectiveness of lie detectors. Even if there were, we can’t assume that evidence from the existing practice involving sex offenders will apply to convicted terrorists.
“Lie detection” is a misnomer
Polygraphs record people’s physiological reactions to questions, such as their blood pressure, heart rate and skin respiration. Some of these can be indicators of deception, but there’s no guarantee of that.
“A specific ‘lie response’ has never been demonstrated, and is unlikely to exist”, according to researchers at Newcastle University.
In other words, they don’t detect lies, but they can be used to detect responses which may be associated with deception.
Calling this “lie detection” also doesn’t get across how these tests actually work in the case of supervising sex offenders. “Most [polygraph tests for sex offenders] are screening in nature”, according to the Newcastle academics.
So rather than people being questioned about a specific incident, they’re questioned about their general behaviours, such as whether they’ve been truthful about their offending history or if they’re keeping to their probation conditions.
Polygraphs can’t detect lies accurately
There isn’t generally very good evidence for the effectiveness of polygraphs at detecting lies, and there’s certainly no evidence they work perfectly.
One important study, conducted by the US National Research Council in 2003, concluded that under normal conditions, “specific incident polygraph tests can discriminate lying from truth telling at rates well above chance, though well below perfection”. This study is old, but is still cited by researchers today.
At the same time, the study cautioned that its findings can’t be used to say how accurate the tests are for screening purposes—more like what’s involved in the probation service. Given the greater ambiguity in screening situations (no longer asking about specific incidents), it said it’s likely that accuracy would be lower than what it found for specific incident tests.
The campaigning charity Sense About Science has a useful guide to what the research in this field says more generally.
There are additional doubts over whether polygraphs can be effective at detecting lies from sex offenders or terrorists. Dr Andrew Balmer, senior lecturer in sociology at the University of Manchester and author of Lie Detection and the Law, told us that: “Terrorists and sex offenders are amongst those least likely to fear the consequences of being found to be lying or to be suspected of lying. Their crimes depend upon a lack of empathy and often on skilled deceptions. These are exactly the circumstances that would render the polygraph less useful.”
The Justice Secretary clarified that he doesn’t see polygraphs as “the be all and end all” and says the government is planning other changes, such as more specialised probation officers and better training.
Even if they can’t detect lies accurately, are they still useful?
Whether or not the tests are completely accurate, there’s an argument that there are other measures of effectiveness, which the Justice Secretary hints at this morning. The Ministry of Justice pointed us to the government’s own evaluation of polygraph tests it had been trialling in England and Wales about a decade ago.
When polygraph testing has been trialled in the UK, it has generally found that offenders are more likely to make what are called “clinically significant disclosures”, such as admitting they’ve breached a condition of their probation.
These may not even happen during the polygraph test itself, as some disclosures were made after the test, in particular if offenders were told they had “failed” some part of the test. Other studies support the idea that testing can draw out more information.
Questions have been raised about how reliable these findings in the UK are though, for instance questioning whether probation officers themselves may be biased in their own assessments of how effective the tests are.
A “bogus pipeline to the truth”?
“Experts have concluded that any effectiveness of lie detectors is probably because people are more likely to tell the truth if they believe that the machine will catch them lying”, according to Sense About Science.
Others argue that, even when people are told the lie detector isn’t 100% accurate, they are still more likely to tell the truth than without a test.
What “works” with sex offenders won’t necessarily work with terrorists
The Justice Secretary uses the example of how polygraphs have been used on sex offenders to justify the government’s plans to roll these tests out to convicted terrorists.
But there’s a limit to how helpful the existing evidence can be.
Dr Balmer told us that: “Existing research on sex offenders and the production of clinically significant disclosures cannot be transferred to the use of the polygraph with terrorists without significant further research.”
He explained that: “Sex offending and terrorism are crimes of a different nature. We have accumulated significant knowledge of sex offenders and their pathologies that lead to and sustain offending behaviours and there is much clinical experience with this population. We have relatively little understanding of terrorists' pathologies and how these lead to or sustain terrorist offences.”