Why do ministers want to use lie detectors against criminals? They do not work

It’s Science Fiction with Stuart Ritchie, a newsletter exclusively for i-subscribers. If you’d like to receive it straight to your inbox every week, sign up here.

Here’s what you may have missed in this week’s King’s Speech: news about lie detectors. As part of a series of measures to combat crime, such as longer prison sentences and greater powers for police to enter buildings where they have found stolen mobile phones, the government also announced it would expand the use of lie detector tests.

There seems to be some confusion as to what exactly this change is. Time reported on Wednesday that “lie detectors are used not only in cases of terrorism and domestic violence.”

However, government briefing notes detailing the policy in the King’s speech simply state that the plans are to “give probation officers more powers… to use lie detector tests on serious terrorists or sex offenders to better manage their risks.” .

Perhaps something broke along the way. But regardless of exactly how they are used, it is surprising—and depressing—that lie detector tests are used in our legal system. not at all – because lie detector tests don’t work.

80-90% exactly?

Here’s what’s encouraging: lie detectors, which take a variety of physiological measurements, often including heart rate, blood pressure, breathing rate and sweat rate; All or part of these indicators are thought to change when someone is lying – are not used in court to determine whether a person is lying. whether the suspect is guilty or not. They are not accepted as evidence in any UK court (a fact worth remembering: if the courts don’t accept them, why should anyone else?).

Instead, they are used against convicted terrorists, sex offenders and domestic abusers to ensure they follow the rules of their license when released from prison. It’s not entirely clear, but it is expected that they will be asked questions about who they spend time with, where they have been and what material they have seen on the Internet, and that they may be sent back if they are caught lying. goes to prison, will be recalled. .

Are we sure that a lie detector can actually tell if one of these criminals is lying? We must trust the evidence, and the British government claims (without citing a source) that polygraph tests are “80 to 90 percent accurate.”

Such figures are often mentioned in discussions about lie detectors and are taken from various scientific studies and reviews. They make the lie detector sound like an almost magical machine, capable of detecting hypocrisy with amazing accuracy. However, there are some serious problems with the types of studies that make a lie detector test so highly accurate.

First, people who participated in a laboratory study of deception, in which they were asked to pretend to commit a crime as part of a game, were unlikely to respond in a lie detector test in the same way as those who were actually accused. crime. In addition, laboratory tests make it possible to be sure who committed the “crime” (since it is not a real crime, but an artificial situation). In real life, the situation is different, where it is much more difficult to verify a criminal case. However, we want to see real lie detector studies that test real suspects.

Therefore, we immediately exclude many studies that were not conducted in real situations where the crime actually occurred.

But the really important thing is related to this. Confessions.

Confession breaker

In a criminal case, often a confession separates the guilty from the innocent, CUT, and there is no external evidence that can make the case conclusive. Imagine testing potential suspects with a lie detector test. The car is buzzing with the suspect: he was caught in a lie! In most cases, you will then ask the suspect to confess.

Let’s pretend they say, “He’s an honest cop.” I did it.” In this case, this is a victory for the lie detector: at least, according to the confession, it managed to identify the right person. We can add to the protocol: the lie detector worked and found the person who committed the crime.

But imagine if they hadn’t confessed. Imagine if none of the suspects confessed. In this case, there is no answer to the question of who committed the crime: a big question mark remains. It’s important to note that you can’t tell whether the lie detector worked or not, because – again, because often all you have to do is confess – They have no evidence that a specific suspect committed the crime..

Such cases are not included in polygraph results because there is no “fundamental truth” about who actually committed the crime.

You can see why this is a problem: the studies will focus on covering only (or most of) the first type of cases where the lie detector apparently “worked”, rather than the second type of cases where it did not. that is, it was the case. T. This amazingly cyclical process is a very plausible explanation for the very high (80-90 percent) polygraph efficiency rates.

A 2019 study put it succinctly: “If lie detector tests did not actually offer anything better than chance accuracy because reliability studies relied on testing confessions.” [results] A researcher chosen in this way may erroneously conclude that the method is virtually infallible.”

Oh my God.

Is it true

Don’t take my word for it. In 2003, the US National Academy of Sciences produced a report, “The Lygraph and the Lie Detector,” which concluded that “nearly a century of research in scientific psychology and physiology provides little evidence to support the proposition that polygraph testing can have extremely far-reaching consequences.” High accuracy.” And since then there has been no revolutionary evidence.

The UK government’s figure of 80 to 90 percent is simply not supported by reliable scientific evidence and we have good reason to believe it is greatly exaggerated. Is the actual number greater than 50 percent? So is a lie detector better than simply flipping a coin to find out if someone is telling the truth or lying? Maybe. It is likely that there is a combination of measurable physiological evidence of deception.

But “somewhere above 50 percent” is not what we want in a criminal justice context, especially if we have fooled ourselves into believing that the actual accuracy is 90 percent and so we may be relying on the machine much more than we should. Not to mention, many people know how a lie detector works and can use so-called “countermeasures” such as certain breathing techniques to fool the test.

A 2003 report states: “Countermeasures pose a potentially serious threat to polygraph testing because all physiological parameters measured by the polygraph can be altered by conscious effort…”

At the very least, we want people using the polygraph to be aware of these and other limitations that reduce its accuracy – and by citing the “80-90 percent” figure, the government has shown that this is not the case.

It is likely that, despite the evidence, many people were misled by the “scientific” nature of the lie detector. You think, “It measures all these different things and produces very convincing-looking graphics.” I’ve seen police use it dozens of times on TV shows! How could this be wrong?

But if we really want “evidence-based policy”, we need much more Look Sciences. We need studies that conclusively demonstrate that this is an accurate tool. We just don’t have it at the moment.

This is not just an “academic” discussion. A lie detector error may allow a terrorist, sex offender, or domestic abuser to continue to live in society despite actually violating their license. Given the risk of relapse, this could be a big mistake.

No lies: the British government should abolish lie detectors completely.

Other things I wrote this week

Besides the hype about lie detectors, another announcement in the King’s Speech was about driverless vehicles. Here’s my article on how they are definitely overrated, but shouldn’t be ruled out just yet.

Science Link of the Week

Remember all the room temperature superconductor drama from a few months ago? Well, there is a whole other drama about room temperature superconductors that is playing out in several scientific journals, including Nature. You can read the latest update here.

It’s Science Fiction with Stuart Ritchie, a newsletter exclusively for i-subscribers. If you’d like to receive it straight to your inbox every week, sign up here.