This is a question we are often asked, and it depends on who you ask.
It also depends on the methodology used for lie detection. Practitioners usually use one or more of the following three methods:
Let’s have a look at these.
Most of you who are familiar with our research know that behavioural data that reveals what someone is thinking and feeling comes from one or more of the six communication channels (see here for peer-reviewed research on this). Most popular lie detection gadgets (like the polygraph and voice stress analyser) only measure one channel – psychophysiology. In other words, they measure a stress response, often related to the emotion of the fear of being caught in a lie. The immediate challenge here is that polygraph subjects can be stressed or fearful due to the mere fact that they are being wired up to a machine. They might also fear being disbelieved when they are telling the truth. So a one in three chance of catching a liar? Not exactly. But this was part of the reason the Jeremy Kyle show was taken off air because it was mis-claiming. ITV’s chief executive, Carolyn McCall, said the show employed trained providers to carry out polygraph tests on individuals who were attempting to settle family disputes in front of a live audience, with an expected accuracy rate of more than 90%. See here.
One of the challenges is that some examiners use tricks in a pretest to “convince the examinee that the polygraph will detect any deception in order to ensure the proper emotional response in a deceptive examinee. If the examinee is already ‘wired up’, the examiner may perform a demonstration of the polygraph’s accuracy, commonly referred to as a stimulation or acquaintance test, which may on occasion involve deceiving the examinee in the process by using marked cards or other ‘tricks’ for this demonstration” (Simon, 1993).
The largest accrediting body for polygraph examiners is the American Polygraph Association (APA). Their Polygraph Examiner’s Course “consists of a minimum of 400 hours… completed in a ten-consecutive-week [up to 17 week] cycle… according to accreditation standards”. See page 7 of this document here.
A little short of the 10,000 hours needed for skill mastery – the idea made popular by the writer Malcolm Gladwell in his book Outliers. This “10,000 hours of practice” rule is based on research by psychologist Anders Ericsson. The rule tells us, a mere 10,000 hours of dedicated practice in your particular field is sufficient to bring out the best in you. Scientists, however, remain sceptical. A recent study, however, by a group of psychologists from five universities, question Gladwell’s wisdom.
So what does the research say about polygraph reliability? The National Research Council examined 57 studies of the polygraph and published their findings in a book entitled “The Polygraph and Lie Detection” and concluded that “overall, the evidence is scanty and scientifically weak” (ibid:212). It also concluded for the specific application of employee screening that polygraph testing “yields an unacceptable choice for… employee security screening between too many loyal employees falsely judged deceptive and too many major security threats left undetected (ibid:7). In a quest for a number, I would estimate [there is a meta-study here for an EIA Group MSc student to get the exact figure!), from scanning the research behind this book, that the average reliability of a polygraph, across applications, is around 62%. Better than chance – as is the public at large, and professionals across the legal domain, who average 54% (Bond & DePaulo, 2006).
The 54% performance figure quoted above in Bond & DePaulo (2006) was cited in a paper that explored the reliability of Judges watching videotaped interviews. They averaged 53%. Bear in mind that these studies are looking at individuals who have not necessarily been formally trained in behaviour analysis or lie detection techniques. I stress this due to the poor researchers who use these statistics to claim that is all a human is capable of. We know differently.
So how much training is needed for mastery?
If we take the 10,000 hours figure, then this would require about 90 minutes of development-focused practice every day for nearly 20 years, or four years full time (48 hours a week!). Ericsson is also on record as emphasising this last point that not just any old practice counts towards the 10,000-hour average. It has to be deliberate, dedicated time spent focusing on improvement.
We conducted a study in Europe where we trained 23 security personnel for 56 hours in how to interpret behaviour and how to engage individuals in forensic conversation/interviews our clients achieved an average of 84.6% reliability in lie detection (see here). Our team are proud to deliver many individuals at 90%+ reliability in lie detection as a result of our training. We also have the world’s first (online) MSc Masters programme in Behaviour Analysis which engages experienced professionals and supports them through a two-year intensive development programme.
It has got to be in training humans in behaviour analysis – as long as it is research-based, focused on 6 communication channels, and involves the engagement of the subject by the trained practitioner.
You don’t need permission to take your eyes and ears into a meeting or on a Zoom call – as you do with a polygraph. It is less intrusive and more practical… consider a suspicion of infidelity by your partner and you request if he/she will strap on a polygraph! The lack of use of technology reduces the stress for the individual you are interested in, so this will lead to fewer false positives (truthful people being disbelieved – we had a 60% reduction in false positives in this study).
Single-channel analysis (like the polygraph, voice stress analyser, facial expressions[only], statement analysis) are risky and unreliable – they are like owning a guitar with only one string – not much use for complex situations/music. Plus the additional channels provide collaboration – we only become interested when we see/hear three indicators across two channels within seven seconds of a stimulus/question. We have to take this approach as many of our clients are operating in life and death contexts.