Lie Detector Testing

Both the conclusion of the BTO, and Reskin, and Likken unanimously and categorically protest against the use of a lie detector for preliminary testing in recruitment. On the other hand, it is a favorite method of many employers, professional lie detector operators and even many government officials, especially those working in intelligence agencies. And despite the fact that the detector is most often used for hiring, research has not yet been carried out on how accurately it detects false candidates. It is not difficult to understand this, for it is not so easy to determine the truth in an on-site investigation. One way to determine this would be to study the situation when all candidates are taken regardless of the test results, and then working supervision is established for them. Another way could be a detailed study of the past work activities of the candidates in order to determine what and how they lied during the self-presentation. But such studies would be very, very expensive. So far, only two studies have been conducted on this topic by analogy, and one of them showed high accuracy of the results of tests on the detector, the other - low ^[155] . In addition, there are too many discrepancies and too many ambiguous in each of these two studies in order to draw any serious conclusions based on them.
It is impossible to judge the accuracy of the results of testing on the detector when applying for a job, based on studies of such tests in criminal practice (see previous subchapter). Here, completely different people pass the tests, the others are both the testers themselves and the techniques they use. In order to get a job, candidates are required to take a test, while criminal suspects have a choice whether to take this test or not, and their refusal does not indicate against them. Reskin says that pre-testing on a lie detector "is mandatory and probably causes a feeling of disturbance, which can greatly affect the accuracy of the results" ^[156] .
What is at stake in both cases is also quite different. The punishment for exposing fraud in pre-testing for admission to work is much less than in criminal cases. And since the stakes are lower, then the fear of exposing the liars here is experienced to a much lesser extent, and it is more difficult to catch them. True, honest and very willing people to get a job may be afraid, they will be assessed incorrectly, but precisely because of this fear they may be incorrectly assessed.
The counter argument set up in such cases by advocates of using a lie detector when applying for a job is that this technique does its job. After such a test, a lot of candidates get admissions for such things they were not aware of before the test. And this is a practical argument. And it does not matter how accurately the detector detects the liars, if the result still identifies those who do not need to be recruited. So, the detector is useful. True, Likken objected, noting that this practicality itself may be unfounded ^[157] .
The number of confessions made to oneself’s detriment may exceed the real state of affairs; Some of these confessions may be fake, made under pressure. In addition, those who made a confession that did not give an opportunity to be accepted could have done so and not from fright in front of the detector. Without accurate research in this area, it is not possible to know how many people who have failed in such tests are in fact honest, nor how many people who have passed this test are robbed of their employers.
Gordon Barland, a psychologist prepared by Reskin, gives completely different conclusions in favor of using a detector. He studied 400 job candidates, such as truck drivers, cashiers, storekeepers, etc., sent by employers to private firms to pass detector tests. Half of the 155 candidates identified as fraudsters admitted this immediately after the test results were announced. At the same time, Barland found that employers, even without waiting for the results, 58% of these liars had already hired. “A lot of employers use detector tests not to decide whether or not to hire a person, but rather to determine which position to determine. For example, if a candidate is recognized as an alcoholic, he will be taken as a docker, and not as a driver ” ^[158] .
Barland correctly indicates that those 78 people who, identified as liars, did not recognize this, are of the greatest interest, for they could simply be victims of the mistakes of unbelief of truth. He says that 66% of them were still employed, but we are not able to find out, took them exactly for the job for which they wanted to get a job, or the test results still prevented them from doing so. Most of the unaccepted and denied their lies, most likely, was rejected at a preliminary interview that occurred before the test on the detector. “Only a very small part (less than 10%) of candidates identified as deceivers, but who did not recognize it, were rejected by potential employers precisely on the basis of tests on the detector” ^[159] .
How to treat these less than 10% and how much damage is caused by them depends on the basic norm of the lie. The base rate is a statistical indicator of validity; for example, the basic rate of perpetrators of criminal suspects and those undergoing detector tests is very high and sometimes reaches 50%. (As a rule, detector tests pass by no means all, but only by a small group of suspects.) Barland’s research reveals that the base lie rate among job candidates is about 20%. Approximately one in five candidates who lie during the tests are not hired.
Even if it is recognized that the detector is more accurate than it actually seems, the basic rate of 20% makes the result too disappointing. Reskin, objecting to the use of a preliminary test on the detector when applying for a job, takes the detector's performance for 90%: “Based on these considerations, preliminary tests on a lie detector when applying for a job would have the following results (based on 1000 candidates): out of 200 the deceiving 180 would be correctly identified as deceivers, 20 the same incorrectly identified as telling the truth; out of 800 truth-speakers, 720 would be correctly identified as truth-telling, and 80 incorrectly identified as deceivers. Of the 260 people identified as deceivers, 80 would actually be true. Thus, of the recognized deceivers, 31% of the true people would turn out to be. This is a very high error rate of disbelief of truth, leading to the rejection of the use of the detector as a basis for decision-making. Such results would not have been in criminal investigations, since the base rate of lies there is 50% or more, and such accuracy of the technique would not lead to such a high level of errors ” ^[160] .
Here the following counterarguments are possible: the estimate of the base lie rate among job candidates of 20% may be underestimated. This result is based on only one study conducted in the state of Utah. It is possible that in states with a smaller number of Mormons, the percentage of liars will be higher. However, even if it reaches 50%, opponents of the preliminary test can say that without evidence of the actual accuracy of the detector, no conclusions can be drawn; and this accuracy is most likely significantly less than 90%.
The accuracy of the detector tests in general does not matter in reality. The test itself, or even the threat of it, forces people to give information that they would never have discovered in other circumstances. The answer again boils down to the need to study the accuracy of the tests on the detector, because without knowing this there is no way to know how many of the unrecognized and employed people will cause many problems to their employers.
The relative advantage of the detector is that it periodically tested by those people who have already been hired. And none of the critics of the preliminary use of the detector deny this benefit.