Report by Andrei Stepanov, a polygraph examiner from St. Petersburg.

[Матушкин Андрей Николаевич]

Report by Andrei Stepanov, a polygraph examiner from St. Petersburg.

The St. Petersburg School of Lie Detection was established on July 30, 1998. From the very first day of its existence, a course was taken to create a highly professional system of individual training for psychophysiological studies using a polygraph ("lie detector") based on the best world practices of the USA, Israel, and Japan.

To organize the educational process, we used the principles of training of the higher technical school of the USSR, worked out for decades - you must agree, it was the strongest in the world. The best system for the development of theoretical material and its practical implementation was taken as the basis - the Plant-VTUZ system, organized in 1930 on the basis of the Leningrad Metal Plant by Henry Ford, the founder of Ford Corporation. In addition, we have fully implemented the training system used at the Higher Courses of the KGB in Minsk in the 80s of the last century.

On the one hand, the use of the polygraph in business and law enforcement bodies have common features, on the other hand, they differ fundamentally. But you probably agree that the mistake of a polygraph examiner in both areas can have equally serious consequences for the subject. Therefore, in my speech I want to dwell on only one direction of our work - the selection of candidates for training.

All our almost 20 years of experience has shown that we can’t take everyone to study, because not everyone has the necessary qualities and abilities to master this complex profession. I will try to explain why. A specialist in instrumental lie detection with the help of a sophisticated methodology penetrates the inner world of the subject and has a manipulative effect on him. According to the results of each study, on the basis of the data obtained, by and large, he decides the fate of a person. So the primary requirement for the candidate appeared - he must first be a decent person. Those. its moral reliability should not cause any doubt. And how to establish it - in only one way - to study all his previous life experience with the so-called. possible manifestations of risk factors and their systematic nature.

Then we realized that not everyone can become a successful specialist - we need certain personal qualities. So there were whole blocks of the study of psychological characteristics, value orientations, communicative, stress-resistant qualities, etc. future candidates for training.

Our profession is very responsible - a mistake can destroy a person’s life. The method is completely probabilistic. Therefore, we came to the conclusion that in order to master this complex methodology, the candidate must have a certain intellect, horizons and special abilities.

And in all these areas criteria were developed for their assessment and decision-making on the suitability of a candidate for training.

The development of criteria for predicting the success of a future polygraph examiner was carried out not from scratch. The fact is that in 1992-1996, as part of the psychological support group for operational activities created in the Leningrad State Security Directorate (this group was the only one in the country), my colleagues and I developed a system for selecting and training negotiators with hostage-takers. And as it turned out, the work of a professional negotiator is very close in all respects to the work of a competent polygraph examiner.

Now let's look at the sequence of studying a candidate, before deciding on his admission to study.

1. Formal biographical criteria.

I do not even consider candidates up to 30 years old for study. 80% of my students are about 40 years old or older. This profession should be handled by a mature person with a good life experience. And the accumulation of this experience, as a rule, occurs by 35-40 years. Candidates without higher education cannot get to me. Moreover, the presence of a diploma - this does not mean how he mastered the material at the institute - this is established during the interview. Moreover, higher technical rather than psychological education is preferred. The experience of family life and the presence of a child are desirable. A person who has a criminal record or was under trial or investigation (depending on the circumstances) also cannot be considered for study. Carefully considered

practical experience in labor collectives. Motivation to work in this profession, a tendency to investigative work is an additional positive factor.

2. The candidate is preparing an expanded autobiography from kindergarten to the present time according to our proposed scenario. Usually it takes 3-4 sheets. At this stage, the verification of personal data and the initial study of the candidate, including by the method of psychological autobiography. Based on the text presented, we look at how the formation of personality, the development of abilities, value orientations, intelligence, and horizons went on. As far as the presentation is properly constructed, the presence of errors, already at this stage we predict the leading personality traits inherent in one or another accentuation. The candidate is checked against possible bases and records.

If the candidate lives in St. Petersburg, then he comes for a two-hour interview. Here - a well-constructed conversation and assessment of the reliability of the information received is fundamental to the adoption of the initial decision. I have to communicate with nonresidents via Skype, which is less preferable than direct communication. Very often after talking with me and fully understanding what he will have to do and what difficulties he will encounter, the applicant refuses the idea of mastering the profession of a polygraph examiner. But the one who is ready to move on - proceeds to the next stage of selection.

3. In-depth psycho-diagnostic testing.

It takes place directly in St. Petersburg within 2 days.

At this stage, the presence of several fundamental psychological blocks is carefully studied to predict success in training and in subsequent work.

A comprehensive study of the inherent candidate leading psychological qualities of the individual. For this, a multilateral personality study is carried out, leading accentuations are identified, the possible presence of psychopathies is determined, the basic properties of the nervous system are studied, etc.

In addition, the candidate must possess:

* A high level of intellectual development, a broad outlook, knowledge in many areas. Cognitive mental processes, long-term and short-term visual and auditory memory, etc. are investigated in detail. In the absence of intellectual abilities, he simply will not be able to master the complex methodology that we teach.

* Developed communicative qualities. The specialist should be able to establish contact with any person who comes for testing and arrange him to himself in the shortest time, to receive information of interest.

* High level of stress resistance. We have a rather stressful and dangerous job. A person coming for examination never wants to tell that he had bad things in life or that he was involved in the events being investigated. For a specialist, in 3 hours it is necessary to study him carefully and double-check whether he told everything, whether he deceived us. The top skill of the polygraph examiner is to obtain recognition of the crime. And this is very stressful. According to the technology developed by us, in 70% of cases we receive such recognition. Another important point: our probabilistic method. We are under stress all the time if we have not received recognition that we have drawn conclusions correctly. And sometimes you don’t have to sleep at night, suffer, but correctly you accused him of not missing the guilty party, etc. This is also stress. In addition, there are inadequate customers, surveyed or those who stand behind them.

* Must have artistic abilities. Without them, I do not take to study. If there are no artistic inclinations, you will never master the methodology of manipulative influence and you will never receive recognition from the subject. None - neither in hiring, nor in investigations. And you will be a simple extras - the operator of the machine, of which 95% of all those working in Russia as "lie detection experts" are machine operators.

T.O. a forecast is made of the psychological and psychological reliability of the candidate.

4. Psychophysiological study on a polygraph.

The absence of serious negative misconduct in the candidate’s biography is a prerequisite. Therefore, the candidate goes through the fourth stage of selection - a psychophysiological study on a polygraph on 30 risk factors. At this stage, a forecast is made of its moral reliability. We check everything from drug use, alcohol abuse, theft, crime, etc. Their seriousness and possible systematic nature are established. If I detect this by any of the “risk factors”, I do not need such a specialist. He crossed the line, which could not be crossed. If he committed regular thefts in life, fraud, etc. “He doesn't interest me either.” Candidates with a criminal record for me at school are closed - I always refuse them. Although, I know the schools in which they then entered and studied. A code of honor must be required.

It should be specially noted that all the above stages of selection are applied by us both to candidates from business and to candidates from law enforcement and administrative bodies. Moreover, the selection stages for candidates for primary basic training and for those already trained in other schools as polygraph examiners, who also take us only for a basic training course, are the same.

Of course, from a business point of view, such a selection is not profitable. Yes, we knowingly suffer very large losses. Although they could follow the principle of all other schools - to take everyone who wants and the income is provided. Our statistics show that out of 10 candidates, all stages of the selection go through no more than 2 (dropout is 80%). There were 11 candidates for the July group - only three passed. Eight people were denied. Now tell me, which school in Russia or in our Eurasian Union can afford to refuse 800 thousand rubles ?! Many who did not get to us go to study in other schools. But I clearly decided for myself that the unique methodology that we teach can only be given in trusted and reliable hands. And only in this way it is possible to combat the ignorance that has developed today in the field of using the polygraph in Russia, and in other CIS countries. With this selection, we implement the universal principle of “do no harm”. Well, not everything should be measured in our modern life only with money, because there are other values that should be present in our life ...

Unfortunately, today in Russia there are difficulties and problems in using the polygraph in law enforcement and personnel work. These problems (some of them were presented by A.P. Pelenitsyn at the conference of polygraph examiners of the Ministry of Internal Affairs of Russia in 2008) include:

1.Psychophysiological problem - seems to be the most important, because without answering the main question - how much the psychophysiological patterns, phenomena and processes underlying the applied testing methods using a polygraph determine their validity, accuracy and reliability - talk about effective methods, organizational and the legal support for using the polygraph just doesn't make sense. High cost of error.

2. Lack of regulatory framework.

3. Organizational problems.

4. Lack of sound research standards.

5. Lack of a single standard for the training of polygraph examiners.

6. Lack of selection system for this profession.

7. Low qualification of polygraph examiners.

8. Problems caused by the low quality of the equipment used (polygraphs).

9. Lack of a quality control system for ongoing research.

10. Lack of a proper court review procedure.

11. Lack of responsibility of polygraph examiners for the conclusions made, etc.

That is why for their solution it is necessary to focus on the "best practices" that can be found in the United States, Israel, and Japan. I will talk about this as part of my speech in a “closed” format.

In almost 20 years of the formation and development of the St. Petersburg School of Lie Detection, it was possible, at least within the framework of one School, to create:

1. A rigorous and scientifically based system for selecting candidates for training in this profession on the basis of a professional specialist’s developed by us profession.

2. A verified scientifically-based modern integrated methodology for organizing and conducting applied psychophysiological studies using a polygraph, which allows minimizing “false negative” and “false positive” errors in this completely probabilistic method. The methodology is a single Standard for graduates of St. Petersburg SDL and in terms of volume totals about 2500 sheets.

3. A system of long-term continuous highly professional training of polygraph examiners: 8 part-time stages for 4 years, followed by annual short-term seminars to maintain professional skills.

4. A system of free remote counseling for graduates on all aspects of conducting polygraph investigations and complex screenings of working personnel.

5. A system of continuous quality control of the work of polygraph examiners and their responsibility for the conclusions made, especially in case of violation of the Research Standard. This system is also a form of training.

This process was completed by us in 2010. We were the only ones in Russia, and, perhaps, we remain the ones who so thoroughly approached the creation of not only selection, but also a scientifically-based training system in the field of polygraph, tied to control and responsibility for the results of graduates. For the invaluable contribution to the creation of the "Integrated Methodology ..." St. Petersburg SDL

It should be noted that only at the end of 2015, the National Lie Detection School (Moscow) finally released the unified, scientifically-based teaching methodology only thanks to its appointment as rector Alexander Borisovich Pelenitsyn, who created this unified standard for NSLD. In addition, the NSLD creates a very interesting portal for remote support of its graduates, unfortunately paid.

The developments of St. Petersburg SDL can underlie the creation of verified, scientifically-based standards for all federal polygraph examiners working in state institutions and polygraph examiners working in business.

All this work was carried out by us in the absence of a regulatory framework, organizational and methodological problems, the lack of a proper judicial review procedure, and the dominance of ignorant low-grade courses, schools, and academies graduating from hundreds of illiterate polygraph examiners. This all created a tough competitive environment. But life has shown that we have chosen the only right way - a well-trained specialist should still be in demand. It is such a professional who will always respect the rights of any citizen of Russia and will not harm him with his illiterate and ignorant actions.

 

[Матушкин Андрей Николаевич]

WHAT YOU NEED TO KNOW ABOUT THE SO CALLED “VOICE DETECTORS”

In the past 35 years, with enviable consistency in the market of lie detection services under various names, but always with a halo of “scientific” and mystery, devices have appeared that position themselves as “stress analyzers in the voice”. Security officers, merchants of various levels, and just ordinary people, not having enough training to appreciate the true essence of these devices, actively buy them under the influence of advertising, so that later they will be convinced of their complete futility and inability to solve problems of interest. On the Internet at https://gdata-python-client.googlecode. ... .18.tar.gz posted an article from the American professional journal Polygraph entitled "Lie Detection", which, in particular, refers to the complete inefficiency of voice analyzers. However, given the importance of the issue, we decided to reveal this problem a little wider.

First of all, in order to better understand what kind of devices they are, we will take a short excursion with comments on the history of its creation. To begin with, two Americans in 1970 — retired colonel Charles R. McQuiston, a former army polygraph operator, and Allan D. Bell, a former army intelligence officer, set up a company in Virginia. Dektor Counterintelligence and Security, Inc. ", which began to produce PSE (Psychological Stress Evaluator) psychological stress analyzers and trained operators to work on them for 5 days. In 1972 they filed an application and on July 20, 1976 received a patent (No. 3,971,034) for their device. Somewhat later, someone Fred Fuller developed a similar device, which was called the Voice Stress Evaluator (VSE). Allegedly, it was based on the same psychophysiological principle as PSE. It was argued that these and similar devices (Mark-II, ESM, Hagoth, etc.), unlike a polygraph, are able to establish insincerity without connecting sensors to the human body, and by measuring changes in the voice due to stress that accompanies false statements. During the verification procedure, reminiscent of a polygraph test, in the sense that the inspector is asked neutral, verification and control questions, his answers are recorded. According to the authors, such a survey could supposedly be carried out even by telephone. For example, an employee of the personnel apparatus could ask a candidate for a job a number of pre-prepared questions, and then analyze the sound recording of the answers to them and get information about the degree of sincerity of the candidate ..

Although McQuiston and Bell argued that the procedure they proposed was unlikely to be reliable if it did not use the format of the method with security questions, many manufacturers and especially consumers of such voice analyzers used them without taking this requirement into account. If, over the past 35 years, conclusive evidence had been obtained that this method has the declared high accuracy and reliability, then it would undoubtedly have long been an effective diagnostic tool in a wide variety of areas of human relations, including business and international negotiations. However, such convincing evidence never appeared.

The essence of the work of stress analyzers in the voice was explained as follows. In the process of analyzing sound recordings of speech utterances using a device allegedly

the acoustic characteristics of the voice, which are not perceived by ear, are caused by stress. In particular, PSE manufacturers claimed that this device measures the low-frequency (about 10 Hz) modulation of the frequency of the fundamental tone of the voice due to muscle tremor. Indeed, the English physiologist Lippold at one time showed that in a state of complete relaxation and rest, skeletal muscles experience myographic tremor (microvibration) with a frequency of about 10 contractions per second. In a state of psychophysiological arousal, this tremor weakens or disappears completely. The essence of the theory behind PSE-type devices is that the throat, pharynx, and respiratory muscles are also (should) be exposed to this tremor, which modulates (overlaps) the speech signal. Moreover, a person’s voice in a calm state will be characterized by the maximum representation of this frequency component, and under stress, its severity will be minimal. However, in direct myographic experiments, this phenomenon was not found in these muscles, just as it was not convincingly technically proved that PSE, VSE and similar devices really measure the degree of low-frequency modulation of a speech signal.

In his advertising review of the practical use of PSE voice analyzers, McQuiston pointed out that in more than 10,000 cases of its use, not a single mistake was made, and in 5574 cases of joint use of PSE and polygraph, a 99.85% coincidence of conclusions was observed. Obviously, such declarative statements have no practical value, since the validity and effectiveness of any diagnostic method or device should be evaluated in accordance with the necessary requirements of a scientific methodology. In addition, the two above figures already contain a serious contradiction, indicating a careless and superficial attitude of the author to the problem of the accuracy and reliability of his brainchild. If the accuracy of PSE really coincides with the accuracy of the polygraph (by 99.85%), and the accuracy of the latter, as you know, is somewhere around 90% (at least, obviously not 100%), then it is not clear how in 10,000 cases not a single erroneous conclusion was obtained when there should have been about 1000 of them.

But this is all reasoning. But what are the results of an independent assessment of the accuracy of these devices. Back in 1974, during special hearings in the US Congress on the use of polygraphs and similar devices, federal departments provided numerous data on the low reliability of PSE analyzers [J. Kubis, Polygraph, 1974, 3, 1-47; G. Barland, J. Acoust. Soc. Amer., 1974, 55, 423] and a legislative decision was passed prohibiting their use in public institutions. Studies of the effectiveness of voice analyzers conducted in the US Army showed that with their help the correct results are obtained only in 1/3 of the cases, that is, the diagnosis is carried out at a level lower than random guessing [G. Barland, Polygraph, 1978, 63, 338-344] (By the way, G. Barland was the first director of the current Polygraph Institute of the United States of America). A similar conclusion about the low effectiveness of PSE and VSE devices in detecting sincerity or insincerity was obtained at the National Security Agency, which, however, recommended trying their capabilities in other areas where it becomes necessary to remotely assess stress. In 1978, during a regular hearing in the United States Congress, Brenner and Brackcomb at the University of Oregon and the Massachusetts Institute of Technology, respectively, introduced new voice analyzer research results [M.Brenner, H. Branscomb, G. Schwartz, Psychophysiology, 1979, 16, 351-357]. Based on their data, a PSE analyzer

indeed, it can reveal the state of stress by changing certain acoustic characteristics of the speech signal, however, these devices are characterized by serious technical shortcomings leading to low reproducibility (reliability) of the results obtained with their help. In particular, their testimony is highly dependent on the quality of the sound recording and the arbitrary change in their voice by the announcer.

F. Horvat, a well-known specialist in the field of polygraph in the USA, conducted a comparative study of the effectiveness of the PSE analyzer and the electrodermal (RGR) reaction in a laboratory test, during which a hidden number was revealed. The diagnostic accuracy using PSE was only 38%, which was even lower than the level of random guessing (equal to about 50%), while the accuracy of revealing a hidden figure based on the electric skin reaction was 92% [F. Horvath, Journal of Applied Psychology, 1978, 63 , 338-344 and 1979, 64, 323-330]. Subsequent studies by J. Kubis also showed that the contact polygraph is most effective in detecting lies, then there is an analysis of signs of non-verbal behavior, and voice analyzers give the worst results. In a controlled experiment, they were practically not able to distinguish between the perpetrators of the theft, its witnesses and those who were completely not involved in it. Suzuki, a Japanese polygraph expert, performed an acoustic analysis of the pitch frequency, intensity and temporal characteristics of 75 answers to relevant questions in real criminal investigations. None of the parameters gave a diagnostic accuracy higher than the level of random guessing, from which it was concluded that the voice analyzers were unreliable and useless in verification activities.

Such is the story, which, unfortunately, has taught the present little. Apparently, these devices will feed those who produce and sell them for a long time to come. Noteworthy in this regard is the statement of the head of the company that produced the Hagoth voice analyzer. To the question of the correspondent of one American newspaper - "So does your device work all the same or not?" he replied - "Of course! If you do not believe me, look at my bank account." After such cynical statements, it becomes clear that all voice analyzers, in whatever types and packaging they are not released, turn out to be akin to such fraudulent brands as Herbalife, Kremlin tablet or Zirconia bracelet. It is impossible to deal with consumer error by turning to their mind - everyone wants to spend money first, and only then make sure that manufacturers and sellers are deceived on their own. That is their right.

If we turn to computer stress analyzers in voice (Computer Voice Stress Analyzers - CVSA), which have recently flooded the low-demanding Russian market, and without going into the essence of these devices, already described above, see at least the instructions attached to them , then immediately affects the low quality of their content. Important psychophysiological ideas are presented primitively, at the level of a housewife and, to a large extent, are simply incorrect. Substantive errors can be found in each paragraph. The material is eclectic, everything interferes in one pile. The proposed methods of using the device are a rough, with poor translation and distortion transfer of a typical examination procedure using a contact polygraph even without trying to adapt it to the conditions of "contactlessness" (which is just one phrase "... the feet should be flat on the floor" (?) found in one of these instructions).

Despite the huge volume of texts on instructions for using voice analyzers, there is practically no information on which

psychophysiological phenomena and patterns laid the foundation for the operation of these devices in addition to the fact that the statements of the very first developers cited above are almost verbatim repeated. But this information is extremely important in order to understand what methodological requirements, conditions and why you need to comply, in order to get the most valid results and control their quality. Unfortunately, this is not. Attempts by today's authors to introduce their own definitions of psychological and psychophysiological concepts and their classification also seem completely unconvincing. But instead of the normal presentation of the substantive part of the method, great attention (the lion's share of texts) is paid to the small and secondary details of working with software that are already obvious to any modern person familiar with the computer. This is a well-known technique that is used to enhance the visibility of the scientific nature and complexity of the problem solved by computer technology.

Since the material presented here is intended for specialists, most of whom are Russian polygraph examiners, you need to know how not individual authors, but the most respected and competent professional organizations in the field of lie detection, treat the problem under consideration. Without hesitation, we can name two of them - the American Polygraph Association (ARA) and the Institute of Polygraph Institute of the United States Department of Defense, whose position forms the principles of using the polygraph and other means of lie detection in all US federal agencies. So, both of these respected organizations are absolutely unequivocal and without any “if”, on the basis of all the studies conducted, recognize, without exception, voice stress analyzers, including the latest version of CVSA, as ineffective and not acceptable for practical use because of their low accuracy and reliability. The leadership of these organizations asks experts in the field of lie detection around the world, using their professional knowledge, to explain the pseudoscientific essence of these analyzers and by all means to oppose the substitution of a complex psychophysiological procedure for detecting lies using such "toys" that are completely unsuitable for these purposes.

And finally, the latest data from the USA at the end of 2005. The attitude towards voice analyzers, both instrument and computer, in serious US circles continues to be negative, especially when it comes to a possible replacement of a polygraph in the future. It is recognized that, first of all, due to aggressive marketing policies and outright lies by manufacturers and sellers of these devices, they are widely used in police practice in most US states. Another reason for the "tumor growth" in their numbers is the complete scientific ignorance of the police and the passive position of the federal government, which, being warned by the American Polygraph Association and the Polygraph Institute of the Ministry of Defense, did not stop this pseudoscientific product in time.

Currently, the use of voice analyzers is prohibited in all federal departments and in 9 out of 50 states, namely:

* Illinois

* Oklahoma

* Michigan

* Texas

* Vermont

* Virginia

* South Carolina

* Kentucky

* North Dakota

A few states are considering banning these devices. In other states, their use is not what is permitted, but simply not regulated in any way. In addition, in accordance with the federal regulations on “Equal Rights in Employment” and “Requirements for Selection Procedures in Employment”, voice analyzers are not allowed to be used in evaluating candidates for work. Thus, voice analyzers are “successfully” used only at the local police station level on their own initiative without any centralized regulation of this process. In addition, to date, President Bush signed the Directive No. 1447 on Security Measures on Aviation Transport, which authorized the Minister of Transport, among other security measures, to use “computer voice analyzers,” for reasons not described below. (But once again, someone really wanted to cash in on the tragedy of September 11th).

Unlike a polygraph, the effectiveness and validity of the use of which is confirmed by numerous adequate scientific studies, although they have not shown its 100 or even 98% accuracy, the effectiveness and validity of voice analyzers has NEVER BEEN SHOWN or PROVEN. On the contrary, the absolute majority of all serious studies (there were several dozen of them over 30 years, including Russian ones) showed that the accuracy of these devices is at the level of RANDOM GUESSING and even worse. The constant statements by manufacturers and sellers of voice analyzers that their products work with 100% accuracy are FALSE, since today it’s clear to the student that it’s not possible to have procedures with 100% accuracy for a person.

According to the results of the fundamental analysis, which the National Research Council conducted in 2003 on the instructions of the US Congress regarding the scientific validity of using the polygraph and other methods for assessing the reliability of testimony, an unambiguous conclusion was made - there is not a single scientific evidence that voice analyzers have at least some scientific justification and can be used in practice. This conclusion once again confirmed the position of the American Polygraph Association (ARA), as well as the authors of this article.

But the most valuable are the numerous practical studies of this issue carried out at the Polygraph Institute of the United States Department of Defense (DoDPI). Almost a dozen such studies were completed in 2002-2003. They all showed with amazing convincingness that all commercial voice analyzers do not have the necessary validity and cannot even count on close proximity to replace a polygraph. It is important to emphasize that no one denies that these devices with one degree or another degree of accuracy are able to determine the presence of severe stress in a person by changes in his voice, but in practical terms they are NOT SUITABLE for instrumental lie detection.

And finally, the last. Last December, the ARA sent an open letter to NITV (National Institute of Truth Verification - the largest provider of computer voice analyzers in the US market; manufactures CVSA - Computerized Voice Stress Analyzer - the most active voice stress analyzer consumed by the US police), in which it officially stated that the high accuracy data of the CVSA analyzer is absolutely not confirmed by independent experimental studies and, in fact, is a direct fraud and scientific fraud. ARA asked for formal open writing

reply to your letter. Until today, no response has been received from NITV. In addition, NITV refused to provide a copy of its CVSA voice analyzer to the National Institute of Justice (NIJ), which intended to conduct its own research on the accuracy of this device and wanted to have a reliable working copy for this, the effectiveness of which would be guaranteed by the manufacturer. The authors of this article fully share this position of the American Polygraph Association and the United States Polygraph Institute of Polygraph. So in the US, little by little fraudsters are pressed.

On March 30, 2006, the "abc-NEWS" agency published a 3-page report according to which it managed to obtain the results of the latest study from the US Department of Defense, which indicates that analyzers widely used by the US police (about 1,500 police stations) and the military police voices work no better than “tossing a coin,” that is, at the level of random guessing. The report provides information that several terrorists were mistakenly released after they passed a test using this device. The report also said that during one of the lawsuits against NITV in connection with fatal errors made as a result of using the CVSA device manufactured by him, one of the top managers of this manufacturer's company admitted that the voice analyzer is not able to distinguish false from truth. Also, with reference to officials, it is alleged that the popularity of this device in the US police is connected, first of all, with the fact that with its help confessions from suspects are easily “knocked out”. That is what they consider to be its main advantage (no psychophysiology and science).

It is further reported that in US courts, judges are increasingly angrily dismissing evidence presented by the police through the use of voice analyzers, calling them scientifically unsubstantiated. An additional scandal is connected with the fact that the head of the NITV, Dr. Charles Humble (see photo), in fact, does not have any doctoral dissertation, and he assigned this degree to himself (he confirmed the fact that he forged a Doctor of Psychology diploma) for "marketing" reasons. Thanks to this reception (just two letters before the surname), the company's profit exceeded a million dollars.

According to Robert Rogalski, US Under Secretary of Defense for Counterintelligence, a large-scale and thorough study conducted just on the instructions of the Pentagon (it’s precisely the results obtained by abc-NEWS reporters) that showed "a complete lack of connection between the voice analyzers and the real presence or lack of lies and stress. "

“We need a more reliable technique than just tossing a coin,” Rogalsky said and expressed concern that several terrorists had been released in Baghdad due to errors in this device.

Currently, the Pentagon has banned the use of voice analyzers in its department.

In 2006, Sujita Bhati and Susan E. Brandon submitted the work, "Researching Technologies Based on Analysis of Stressful Voice Changes to Detect Deception," to the US Department of Defense. They studied 24 empirical studies that were conducted over a period of more than 30 years, and concluded that these studies did not confirm the correctness or reliability of technologies based on the analysis of stress changes in the voice to detect fraud. For these studies, Multilayer Voice Stress Change Analysis (LVA), Computerized Voice Stress Change Analysis (CVSA), Psychological Stress Assessment (PSE), Vericator, Diogenes, Mark II Voice Analyzer and Voice Stress Analysis (VSA) were used.

The following is a partial list of studies in the field of stressful voice changes in detecting deception and their results:

P. Damphouse, M. Apcherch, "Assessment of the validity of tools for analyzing stressful voice changes in a prison", Oklahoma Department of Drug Addiction and Mental Disorders, 2007. Conclusion: LVA and CVSA tests have low reliability in detecting fraud - the basic accuracy was 68 %

G. Hollien, Voice stress analysis analyzer performance assessment, University of Florida, 2006. Conclusion: LVA and CVSA tests are not reliable in detecting fraud. LVA accuracy is below random; CVSA accuracy was slightly higher than random.

Cassidy, Assessment of problem-solving protocols in identifying fraud through voice analysis, 2006. Conclusion: CVSA accuracy is less than random.

M.J. Gianniro, V.L. Sestaro, "The effectiveness of fraud detection studies using the Computerized Voice Stress Analysis Analyzer (DoDPI95-P-0016). Department of Defense Polygraph Examiners Institute, Fort McClellan, Alabama, 1996. DTIC AD Number A318986. Conclusion: CVSA Deception Detection Accuracy Level as a device for measuring the level of stress changes in the voice corresponded to random.

V.L. Sestaro, Comparison between the accuracy of decisions made using the polygraph and the Computer Analyzer for Stressful Voice Changes (CVSA) in Insecurity (DoDPI95-R-0002). Department of Defense Polygraph Examiners, Fort McClellan, Alabama, 1995. Conclusion: CVSA accuracy slightly exceeded random.

D. O'Hire, M.J. Cody, S. Wang, E.I. Chao, Stressful voice changes and the detection of deception in the Chinese, "Communication Quarterly" 38 (2) (spring) (1990) 158ff. Conclusion: Partial reproduction of a previous study. Scores for voice were not associated with deception.

H. Hollien, L. Gayson, J.W. Hicks J., An Analysis of Stressful Voice Changes and Lie Detection, Journal of Investigative Sciences 32 (2) (1987) 405–418. Conclusions: The level of stress detection is random. The level of lie detection is random.

R.F. Valn, R.G. Downey, “An Analysis of Stressful Voice Changes: Using Telephone Records,” Journal of Business and Psychology, 1 (4) (1987) 379–389. Conclusion: The methodology for determining stress changes in the voice has not shown sufficient reliability to justify its use as a selection procedure for hiring.

D. O'Hire, M.J. Cody, R.R. Benke, Fear of communication and stressful voice changes as evidence of fraud, Western Journal of Speech Communication, 49 (1985) 286–300. Conclusions: In only one subgroup, the detection coefficient was much better than random, and the difference was the smallest. The use of controversial statistical methods in this study suggests that modest positive results will not be reproduced in another study. See the following quote.

B.F. Fuller, Reliability and Reliability of Interval Measure of Voice Stress, Psychological Medicine 14 (1) (1984) 159–166. Conclusion: The reliability of measurements of the level of voice stress was low.

H.V. Timm, Efficiency of an instrument for assessing the level of psychological stress in detecting deception, Journal of Theory of Police Service and Administration 11 (1) (1983) 62–68. Conclusion: The level of lie detection is random.

M. Brenner, H. Branscombe, G.E. Schwartz, Instrument for assessing psychological stress: two tests of voice measurement, Psychophysiology 16 (4) (1979) 351–357. Conclusion: "The reliability of the analysis for the practical detection of lies is debatable."

DoDPI Research Staff, J.L. Meyerhoff, G.A. Saviolakis, M.L. Koenig, D.L. Yurik (in the press). Physiological and biochemical measurements of stress in comparison with the analysis of stress changes in the voice using the Computer Voice Stress Changes Analyzer (CVSA) (DoDPI01-R-0001). Polygraph Institute of the Ministry of Defense. Conclusion: A direct CVSA test for medical stress markers (blood pressure, plasma corticotropin, salivary cortisone) showed that CVSA experts could not determine known stress. This project was implemented in conjunction with the Walter Reed Military Research Institute.

F.S., Horvath, Effect of various motivational instructions on the detection of deception by an instrument for assessing psychological stress and galvanic skin reactions, Journal of Applied Psychology 64 (3) (June) (1979) 323–330. Conclusion: Stressful changes in the voice did not lead to the detection of fraud in a degree exceeding random.

B.E. Lynch, D.R. Henry, Reliability Study for Psychological Stress Assessment, Canadian Journal of Behavioral Sciences 11 (1) (1979) 89–94. Conclusion: Stress detection with voice at random level.

F.S., Horvath, Experimental Comparison of the Instrument for Assessing Psychological Stress and the Galvanic Reaction of the Skin upon Detection of Lies, Journal of Applied Psychology 63 (3) (June) (1978) 338-344. Conclusion: The level of lie detection is random.

J.F. Kubis, “Comparison of Voice and Polygraph Analysis as Procedures for Lie Detection (Technical Report No. LWL-CR-03B70, Contract DAAD05-72-C-0217), US Army Ground Combat Laboratory, Aberdeen Proving Ground, Maryland , 1973. Conclusion: The level of lie detection by voice analysis is random.

A. Suzuki, S. Watanabi, I. Takeno, T. Kosugi, T. Kasuya, “Possibility of detecting fraud by voice analysis”, Reports of the State Research Institute of Police Sciences 26 (1) (February) (1973) 62–66. Conclusion: Voice changes were not reliable and not effective.

Probably, such conclusions were to be expected. Stressful voice changes are a tool that defines one parameter, while a polygraph defines three parameters. If we looked at any individual component of the polygraph, we would find that its accuracy is not comparable with the accuracy achieved through the synergy of three parameters.

There are some law enforcement agencies that use stress analysis of voice changes as a tool to determine the truth. Unfortunately, this is more likely the wrong financial decision than an informed choice. Learning to identify stressful voice changes takes 5 days. To learn how to use a polygraph, you need at least 10 weeks. In addition, these authorities report that the use of stress analysis of voice changes has led to an increase in the amount of information and recognition. We are convinced that this statement is true. However, if this is the only goal, we suggest that they attach the wires from the suspect to the copier, inside which there will be a sheet with the inscription "This is FALSE!" The resulting recognition will be the same. This refers to the usefulness of the test, and not to the accuracy. An interview was held with Charles Humble, president of the National Institute for Truth Verification, which produces CVSA, in one of the central television shows [2]. During the interview, Humble admitted that there was no scientific study that would confirm the credibility of the CVSA, only practical use reports from law enforcement agencies using this device. He also stated that he received a degree in psychology after 6 hours of correspondent instruction in religion at a school located in the same building as his company. Unfortunately, despite the fact that there is practical value in using devices for analyzing stressful voice changes to obtain information and confessions, if the future of a suspect or job applicant is decided on the basis of the accuracy of the data collected from stressful voice changes, then this is a hoax!

As an example, we can cite the K-Factor (HR) personnel verification system distributed in St. Petersburg by the Israeli company Nemesysco based on the use of LVA (multi-level voice analysis) technology. Based on the analysis of the truth (falsity) of the information reported by the tested, the system determines the presence of risk factors in its behavior and gives a forecast of their possible manifestation in the future. According to the developers, K-Factor determines the risk on the following topics:

- honesty and loyalty to the company;

- attitude to confidential information and secrecy;

- honesty and professional honesty;

- teamwork;

- theft from the employer;

- tendency to bribery;

- fraud and deceit;

- addiction to drug use;

- attitude to drinking and drinking;

- attitude to gambling, etc.

The test for each topic contains six to eight questions. Asked questions relate to the tested person’s attitude to the topic under study, his specific actions or the possibility of taking actions within the framework of the topic under consideration. In this case, the test person is given complete freedom in answering the questions presented. No special training for personnel working with the system is carried out. Only 15 minutes carried out

briefing. A report with the results is obtained immediately after testing, after processing by a computer and does not require additional decryption. It contains risk assessments in percent (low, medium, high) on test topics. For example:

- declared qualifications - the risk is low (i.e., during the test, no lies, insecurity in answering questions related to education, previous work experience, or the assessment of professional qualifications);

- loyalty to the company - the risk is high (i.e. during the test, when answering questions related to loyalty, the parameters characterizing the reaction reached threshold values);

- a tendency to bribery and protectionism - the risk is high (i.e., during the test, the parameters characterizing the reactions to answers to questions related to the topic of bribery reached threshold values).

The developers claim that the advantages of K-Factor over other systems (intelligence tests, professional suitability tests, personality questionnaires) used in personnel testing are as follows:

- there is no way to prepare for the test when re-testing,

- there are no special trainings to prepare for testing,

- the test person cannot control the parameters for evaluating the test results,

- there is no limited set of answers to questions,

- the influence of the employee conducting the test on the test person is excluded, as well as the possible subjectivity in the expert’s assessments is excluded,

- message to the test takers of false information does not affect the result.

In their opinion, K-Factor is spared the shortcomings inherent in other testing systems, since it assesses not the substantive side of the answer, but the psycho-emotional reactions during the test, which are reflected in his speech.

However, the real benefits of conducting such tests in personnel work may be insignificant. The integrated risk assessment issued on this topic is not clear on what is based, either on hiding information about already completed or planned actions, or on the increased importance of the topic as such for this subject. The fact is that the significance of the topic can be associated not only with a tendency towards violations on the part of those tested in this area, but also with many other reasons, which will lead to the appearance of psychoemotional reactions during testing. But since it is not clear from the assessment in connection with which question these reactions arose, there is accordingly no way to clarify this. Thus, the likelihood of false positive conclusions is increased, and the technique as a whole can suffer from an accusatory bias.

Since the content of the answers is not analyzed in any way, the testee may not speak on the required topic, which accordingly will not lead to the appearance of psychoemotional reactions related to this topic.

In general, after conducting such testing, managers have a set of risks that are not related to any specific information. While, when conducting a screening test on a polygraph, there is a conclusion on similar risk factors based on specific information received from the test, as well as hiding information on very specific issues. This raises the question of responsibility, as well as legality, when deciding on the basis of data from such testing, since it is not clear with what particular points in the test person’s answers the identified level of risk is associated.

We invited the representatives of this company to conduct a comparative analysis of the efficiency of the contact polygraph and the technology they offer. However, so far our proposal has not received an answer.

We also carried out a three-year work to compare the reliability of the information obtained using a contact polygraph and a more serious speech stress analyzer, implemented by one of Moscow commercial firms. For ethical reasons, I do not indicate the brand of the device and the company. The coincidence of the information obtained on these devices was 33%. It should be noted that the information obtained using the contact polygraph was confirmed by the respondent’s confessions on one or another of the identified “risk factors”, and the data on the stress analyzer did not show the presence of reactions that indicate lies, and with detailed answers to these topics.

Based on the foregoing, we want to advise everyone to think about it before spending money and considerable money on acquiring so-well-advertised speech “lie detectors”. Or buy this device, work on it and make sure of the above by us or refute us.

A. A. Stepanov, Head of St. Petersburg School of Lie Detection (St. Petersburg)
A. B. Pelenitsyn, Rector of ANO DPO “CPP”, Candidate of Biological Sciences (Moscow)

 

[Плотников Юрий Михайлович]

Thank!

 

[Детектив-Молдова]

The voluminous report at the 7th Conference was accompanied by valuable video material!
Many thanks to the author!

 

[Виктор Сергеевич]

Andrei Alexandrovich is always an interesting speaker.

 

[Сергей Анатольевич]

Andrei Alexandrovich is always an interesting speaker.

I support.

 

[Дмитрий Викторович]

Andrei Alexandrovich is always an interesting speaker.

I support.

That's for sure!

 

[Николай Николаевич]

Thank.