Are Lie Detectors a Scam?

 by Natalie Moras

Polygraph tests have been used for decades as a tool to detect deception in various settings, from criminal investigations to pre-employment screenings. To many, polygraphing is mostly associated with entertainment - in TV shows (like Jeremy Kyle, Dr Phil and Jerry Springer) and more recently in youtube videos of various influencers and celebrities (for example the Vanity Fair lie detector series). However, the validity of these tests is not nearly as talked about and this leads to the misconception that lie detector tests are 100% accurate. 

Lie detectors are generally depicted as being a source of objective truths, scientific and accurate. However, according to neurologists “these tests are utter pseudoscientific nonsense”. In fact, scientific evidence refutes the efficacy of not just polygraph tests but the newer functional MRI ‘lie detector’ tests too.

Polygraph tests are not necessarily ‘lie detectors’ and merely measured responses to stress. The traditional polygraph (psychological detection of deception [PDD]) is used in law enforcement, at pre-employment interviews and local police departments. A polygraph (meaning many graphs) measures fluctuations in blood pressure, heart rate, respiratory rate and skin electrical conductivity. When a subject is hooked up to the machine and lies when a key question is asked, the subject is likely to become nervous. This results in the activation of the sympathetic nervous system, to change the aforementioned physiological measures: blood pressure can increase, heart rate can increase, respiration can change in various ways (like temporarily pausing or having an increase in the expiratory phase) and an increase in sweat can cause higher skin conductivity due to the salt and water in it. Here, it is understandable that the polygraph measures nervousness, but this nervousness is not necessarily attributed to every subject lying, for every question they are asked resulting in a positive lying result - nervousness can originate from the situation itself as opposed to the pressures of lying or the severity of a question even if the subject is telling the truth. Even the overall room temperature or the phrasing and tone of voice of the interviewer could influence the results dramatically. There is a known problem in blood pressure readings (white coat hypertension) where some patients have consistently higher blood pressure readings whenever they are tested in a medical setting compared to when they are tested at home. This is typically because the patients find visiting doctors scary and the same thing could be happening in a lie detector test - particularly in police interrogations by fearing the police even if the subject has never had a negative experience with them.

To measure the responses of the subject, a polygraph machine needs multiple components to measure respiration, perspiration, heart rate and blood pressure. The pneumograph measures the subject’s breathing and there are 2 major variants of these. One uses 2 sets of rubber tubes that circle the chest and abdomen that stretch when the subject breathes to measure respiration. The other pneumograph involves 2 pairs of electrodes on either side of the chest and then runs a high frequency low amplitude current through the subject's chest cavity. Due to air being a better electrical insulator than flesh, the chest will block more of the current as the lungs expand. To measure the sweat of the subject during the test, a galvanograph is attached to the subject’s fingertips. This uses electrical current to measure the conductivity of the skin (which is increased with more sweat). The cardiosphygmograph uses an inflatable blood pressure cuff to amplify the sound of the subject’s blood flow, and therefore heart beat, to record the blood pressure and heart rate of the subject. The Data Acquisition System (DAS) collects data from each of the subsystems and converts it all into a readable format. Hence, all the data is quantifiable but not necessarily accurate in accessing lying.

The examiner usually interviews the subject in general to establish a baseline of the polygraph and to generate critical questions in the determination of an overall pass or fail. The baseline is established using a control question test (constructed by the examiner) to establish known lies and truths that are unrelated to the critical questions before addressing them to justify judgements made by the examiner. This involves a Stim test, where the examiner asks the subject to lie deliberately, to then show the measured changes on the polygraph when one lies. This is done to convince those involved that the polygraph is effective in detecting lies. Subjects can therefore be encouraged to confess voluntarily to questions that they might not necessarily have done before and make them more nervous if they intend to lie to generate a bigger (and thus more easily detectable) lie response. They may also conduct a guilty knowledge test (particularly in more professional settings as opposed to entertainment ones). For instance, in order to investigate the involvement of a suspect in a crime they show a range of objects to the suspected criminal and focus on their response to the object that was a weapon found at the crime scene and judge whether or not that response is more severe than the responses on the control question test.

Polygraph tests have been shown to be highly inaccurate and even biased against the innocent - leading to false convictions in areas where they are still admissible in court. They could induce false confessions - particularly when the subject is highly pressured into a confession by increasing their detectable nervousness as previously mentioned. Polygraphic evidence is mostly inadmissible in courts and is also not to be used on current, past or future employees by employers due to the lack of being scientifically accurate. (Whilst there are exceptions to this in government scenarios, polygraphs should never be used alone as evidence for any point.)

In a meta analysis of 34 studies and 3099 polygraph exams, under the ideal circumstances of experienced examiners and examinees untrained in countermeasures with no incentive to beat the test, the median accuracy was only 86%. The probability of correctly identifying a subject lying ranges from 55% to 95% but more astoundingly the false positive rate in individuals ranged from 0% up to 65%. Additionally, a small meta analysis of 7 studies in law enforcement reported a mean accuracy of 89% - which is better but could result in more severe consequences than other examinations. 

They are also subject to countermeasures to “beat” the test. The subject may try to show a bigger response to the control questions so that responses to critical questions are viewed as less severe. There are multiple famous examples of people who beat lie detector tests including Ana Belen Montes (a former senior analyst at the Defence Intelligence Agency in America who, for 17 years, spied on behalf of the Cuban government), Gary Ridgeway (also known as the Green River Killer who was released after passing a polygraph only to later be convicted of murdering 49 women in Washington and confessing to even more) and Aldrich Ames (a Russian mole who infiltrated the CIA and during his time there passed two lie detector tests). The test works on the assumption that a subject's reactions are involuntary, which is not entirely true. 

Douglas Williams administered thousands of polygraph examinations between 1972 and 1979 for the Oklahoma City Police Department and “never once believed that the polygraph was capable of detecting deception”. He described fear as always being around his workspace and how it could be sensed. He stated that “The polygraph is not a test. It’s an intense interrogation and the only part the polygraph plays is to frighten and intimidate a person into making a confession or an admission.” He states that it doesn’t record truth or deception and goes on to say that there is no such thing as a “lying” reaction. He states that it solely detects nervousness and the reaction that brands a subject as a liar only indicates deception “about 50% of the time” - claiming flipping a coin has the same degree of accuracy. Williams asserts that “anybody can pass any lie detector test” by simply duplicating the physiological response to fear at the appropriate times. The more developed a subject's conscience, as opposed to a hardened conscience, the less likely a subject is to pass a polygraph. 

In 1979, Williams started teaching people how to beat polygraph exams. There are 2 types of questions: relevant and control. The control questions are interspersed with the relevant questions in the interview and subjects should show a reaction to the control questions and no reaction to the relevant questions. This is because these reactions are compared in the judging process. Williams states that during the relevant questions one should picture a calm scenario (like “laying on the beach and watching the waves gently roll into the shore”) and in the control questions the subject should picture a stressful scenario (like “falling off of a building”). This led to him, in 2012 and 2013, being the target of a federal sting operation wherein 2 undercover Customs and Border Protection agents seeked his help in using his techniques to beat polygraph tests, and even announced to Williams that their intentions were to use his methods to lie in polygraph tests to conceal crimes. He complied with their requests to teach them which eventually resulted in his conviction in 2015 on multiple accounts of mail fraud and witness tampering. 

fMRI lie detection is thought to be a better model by scanning the brain and seeing which areas of the brain are active when the critical question is asked to determine lying. If these areas light up on the scan, the subject must be lying by the inability to control one’s own thoughts, according to this approach. Whilst there is positive evidence for this, various authors have been selective in their publications and results are more likely to be affected by other factors such as individual nervousness. Studies linking deception to specified regions of the brain all contradict each other by identifying many different regions of the brain responsible for deception - making it hard to attribute any of them to the action of lying. In a study on mock crime (where people were told to commit a crime (such as theft) intentionally) the fMRI detected 9/9 liars but also only 5/15 were correctly identified as innocent - meaning that 10 people were falsely accused of committing the crime. This renders the test practically useless by being so biased against the innocent. 

The lack of accuracy in the lie detector tests has not stopped the use of them by various law enforcements from using polygraphs in hiring employees or researchers in finding an equivalent through microexpression tracking electrodes, eye tracking cameras and facial recognition technology that is adapted to spot a lie at a distance via security cameras - but the long term results of these systems are still being analysed. 

When considering lie detector invalidity, it raises the question of why they were even allowed to be widely used in the first place. In fact, John Larson himself regrets his invention of the polygraph, saying, near the end of his life, that it was “a Frankenstein’s monster, which I have spent 40 years combating”. In 1921, this medical student and police officer studying at the University of California at Berkeley was asked by the Chief of Police there to build a polygraph suitable for police work. Federal agencies quickly implemented this device in their work until, in 1923, it was ruled that polygraph test results could not be submitted as evidence in federal cases, due to a lack of verification of the machine’s accuracy among scientists. This didn’t stop their usage though. In fact, in 1998 (United States v Scheffer) a ruling allowed states in the US to decide for themselves whether or not polygraphic evidence is admissible in court. 

In recent times, polygraph tests have been used as an entertainment gimmick. The most frequently used polygraph examiner in media (according to his website making more than 1000 media appearances - from copious numbers of youtube videos to television shows (such as Ru Paul’s Drag Races, Flavor of Love, Dr Phil, For the Love of Ray J and the Late Late Show with James Cordon) is John Grogan. He has been outed as being unreliable by many, including Ray J after seeing the Kardashian interview on the Late Late Show with James Cordon (with the background of having worked with Grogan before) and Trisha Paytas (who alleges that he simply asks those involved for the result he should give them to each question). This criticism is not limited to people in the media though (who people may critique the bias, with the intention of defending themselves or spreading gossip,), through accusations of him running a polygraph training course with inaccurate information, intimidating and harassing people who did take this course when they attempted to advertise their own services, that he is not technically certified as a polygraph examiner (by not completing at least 8 weeks of continuous, full time, in house training and no less than 400 hours of instruction), allegedly telling people who took his course to charge a standard fee of $395 (despite his website boasting about his low price of $145), encouraging at home testing and immediately feeding back results despite the dangers of this in situations of irate and deceptive examinees (contradicting his website), allowing those who take his test to move around freely and be as talkative as they want despite this being incorrect polygraph procedure and allegedly buying a large number of website domains to redirect people to his website to create a monopoly over polygraphing. Even on his own website, he states that in his entertainment appearances “some are real, some are not and some are a little bit of both … plus, many have ‘scripted’, ‘predetermined’, ‘entertainment’, or ‘known solution’ results”. Hence, the most seen polygrapher is not credible at all, especially for his examinations in the media. 

These inaccurate media polygraph tests have more severe consequences than people realise. On the Jeremy Kyle Show, polygraphs are a main feature in ‘solving’ disputes of members of the public who make appearances on the show. One participant, Steve Dymond, was found dead after failing a lie-detector test that said that he was cheating on his fiancée. Considering that these lie detectors are so inaccurate, it is shocking how they were permitted to be used in shows like this when dealing with large, personal problems - and how they justify doing so for the sake of entertainment. The results make magnitudinous changes to the lives of the people involved and it is not even made clear enough the degree of inaccuracy with the tests - causing people to particularly detach from those they were connected to that were innocent of the accusations made towards them (due to the large frequency in false positive lying results). The audience in these shows boo the potentially lying contestants when they fail polygraphs, with the host hurling insults at them too, destroying their image publicly also and not leaving spectators any space to consider the validity of the results and simply deeming the subject as a liar. The Jeremy Kyle show was discontinued after this event and an investigation surrounding those on the show and their results was started but it’s shocking to see that this show went on for 14 years without anyone stopping it.

The main conclusion is that polygraphs and other ‘lie detecting’ systems don’t work and that they should stop being encouraged before more individuals face the harmful consequences of them when they are wrong.