General Questions
In October 2009, Credibility Assessment Technologies LLC (CAT) was formed to bring a new lie detection technology, based on an ocular-motor deception test (ODT), to the market. In September 2013 the technology was given the brand name EyeDetect. On December 12, 2013, the company was officially renamed Converus, Inc. The name Converus is derived from two Latin words: con (meaning “with”) and verus (meaning “truth”).The company is currently headquartered in Lehi, Utah, USA.
EyeDetect uses a statistical method to analyze independent ocular data. The data is gathered and tracked by an infrared camera that observes changes in eye behavior during a question & answer session with an examinee. The test is administered by a computer. Resulting from the analysis of the examinee’s responses and the ocular data, a “binary” outcome is derived called the Converus Credibility Score.
Some of the independent variables used by EyeDetect are pupil dilation, response accuracy, response time, gaze fixation, blink rates, reading behavior and other variables. The object of this logistic regression analysis of various independent variables is to obtain a biologically reasonable answer to describe the binary (two) characteristics in question: credibility or deception.
The score used by EyeDetect intends to “maximize the likelihood” of the categorization of deception or credibility. When a person obtains a Converus Credibility Score between 1 and 49, they are categorized as deceptive. When a person obtains a score between 50 and 99, they are categorized as credible. And because EyeDetect uses a statistical formula, there is a range of credibility scores. The closer the Credibility score is to 1, the likelihood of deception is maximized. Conversely, the closer the score is to 99, the likelihood of credibility is maximized. Thus, if a person obtains a Converus Credibility Score of 51, 52, 53, etc., the probability of credibility is minimal.
In 2006, after the Converus Science Team completed substantial testing of this concept of monitoring eye behavior to detect deception, a University of Utah psychology graduate student working with the science team published its findings. The Osher Dissertation documented the first lab study that demonstrated the effectiveness of an ocular-motor deception test (ODT). A second formal scientific study in 2008 confirmed the effectiveness of the ODT technology, and its results were published in the Webb Dissertation in August of that year. In 2012, additional field studies were conducted. The results were peer reviewed by other scientists and professors and published on April 30 of that year in the Journal of Experimental Psychology: Applied.
In 2002, John C. Kircher and his colleague, Doug Hacker, an educational psychologist with expertise in the psychology of reading, were driving to Seattle to climb Mt. Rainier. En route, they wondered if changes in eye movements and pupil size while reading and answering questions about a crime would reveal deception. They asked themselves, “Would changes in cognitive load affect the eye in such a way that we can capture those changes and be as accurate as the polygraph in predicting whether or not someone is being deceptive?” Thus the idea for an ocular-motor deception test (ODT) was born — later to be branded as EyeDetect.
In 2003, Professors Kircher and Hacker, along with cognitive scientists Anne Cook and Dan Woltz, formed the Converus Science Team. They began working together to produce and validate an ODT solution for deception detection. David C. Raskin joined this science team in 2009.
It should be noted Professors Kircher and Raskin are internationally-known and highly respected scientists in the polygraph community. They frequently consult and lecture on this subject, as well as provide guidance to the polygraph community, government agencies, legislatures, and the courts. They first published research on polygraph technology in the 1970s and then spent 10 years developing the software and hardware for the world’s first computerized polygraph system, which they marketed in 1991. They recognized the need to find new lie detection methods that could complement the polygraph because the polygraph measures emotional response, not concealed knowledge.
In April 2014, after more than 10 years fine-tuning this technology, these five dedicated scientist were finally able to see their years of work bear fruit when EyeDetect was released to the market.
88% accuracy in 30 minutes for screening tests.
90% accuracy in 15 minutes for single-issue testing.
EyeDetect is the first ocular-motor deception test (ODT) solution for deception detection. EyeDetect is an accurate, cost-effective, efficient, secure, and noninvasive method for businesses to manage risk and ensure workplace integrity, and for law enforcement agencies and governments to detect deception. In September 2013 the technology was given the brand name EyeDetect.
Usage & Applications
EyeDetect is designed to detect lies and deception in individuals. A guilty person that responds honestly to questions about an infraction or crime will not exhibit the same cognitive changes that a guilty person exhibits who attempts to conceal their deception. If a person responds truthfully about using drugs during an EyeDetect test, for example, they may or may not exhibit cognitive change. Theoretically, you can’t predict in advance if the person will fail or pass in this scenario.
EyeDetect is designed to detect deception in individuals. If a person confesses before an EyeDetect test, the exam is no longer necessary for the purpose of detecting deception. If the person were to proceed with the exam and answer the questions truthfully, admitting their guilt, they would not exhibit the same cognitive change that a guilty person exhibits who attempts to conceal their deception. Therefore, there is no real need to conduct an EyeDetect test on a person that has confessed.
Any substance or medication that affects reaction time, eye movement or pupil dilation in general will affect comprehension, eye behavior and cognitive load. Test results in an EyeDetect exam will be compromised in this situation.
EyeDetect is not designed for an event-specific line of questioning, such as is used in a police interrogation. EyeDetect is designed to pre-screen job candidates and to conduct periodic evaluations of current employees in all types of organizations. Tests measure an applicant’s or employee’s participation in theft, fraud, money laundering, bribes, drug use, identity theft, violent crimes and receipt of inappropriate benefits at work.
In the U.S., EyeDetect can be used for screening or evaluating federal, state or local government employees. This includes law enforcement and national security agencies.
Outside the U.S., EyeDetect is used in all types of organizations to pre-screen job candidates and to conduct periodic evaluations of current employees. Tests measure an applicant’s or employee’s participation in theft, fraud, money laundering, bribes, drug use, identity theft, violent crimes, and receipt of inappropriate benefits at work.
There is no EyeDetect test to measure fidelity or loyalty. EyeDetect is used to pre-screen job candidates and to conduct periodic evaluations of current employees. Tests measure participation in theft, fraud, money laundering, bribes, drug use, identity theft, violent crimes, and receipt of inappropriate benefits at work.
EyeDetect is used to pre-screen job candidates and conduct periodic evaluations of current employees. Tests measure participation in theft, fraud, money laundering, bribes, drug use, identity theft, violent crimes, and receipt of inappropriate benefits at work.
EyeDetect Functionality
Polygraph machines sense changes in glucocorticoid (hormone) levels, which initiate chemical reactions such as changes in respiration, heart rate and skin conductance (a moisture buildup under skin). Thus, a polygraph measures physiological changes when under the stress of questioning. That questioning is intended to determine whether a person is lying or being truthful. EyeDetect measures changes in pupil size and eye movements that reflect the changes in their brain activity while a person reads and responds to a questionnaire.
For the polygraph, the primary testing premise is that a deceptive person will show stronger emotional reactions to questions about topics for which they are deceptive. For EyeDetect, the primary testing premise is that a deceptive person will show an increase in cognitive load when questioned about topics for which they are deceptive.
Each test consists of an introduction, a practice test, and then five sessions of 48 true or false statements. Questions related to two issues are asked, as well as neutral questions. The same questions are repeated in a randomized order during the five sessions, and these sessions are separated by four sessions of alpha arithmetic-based questions.
Simply stated, polygraph tests that cover multiple issues (not related to a specific incident) may show substantially reduced accuracy and efficiency. The risk of error ¬increases to the extent that the test includes a number of separate issues that can be independently answered truthfully or deceptively. Such tests are problematic when the individual is truthful to some relevant questions and deceptive to others.
A field study done for the U.S. Secret Service (Source: Raskin, Kircher, Honts, & Horowitz, 1988) revealed that accuracy declines when the examinee answers one or more questions truthfully and one or more deceptively for tests with multiple issues. This data argues against conducting a polygraph test that covers a range of issues, such as drug use, theft, bribery, or other criminal activity, etc.
The field study cited above looked at individuals that were either truthful or deceptive to all relevant questions vs. those in which the person was truthful to at least one relevant question and deceptive to at least one other relevant question. For those subjects, almost half of the outcomes were inconclusive (49%) and only 74.5% of decisions were correct.
The American Polygraph Association (APA) cites a study that indicates levels of accuracy as high as 93%. That study was based on polygraph tests done for “event-specific diagnostic examinations used for evidentiary purposes.” In other words, that study cites data from event-specific tests that are part of an interview, based on evidence; in other words, they involve one event with various, specific issues related to that event.
EyeDetect uses a test format known as the Relevant Comparison Test (RCT), which was developed for an automated polygraph screening system; the RCT covers two unrelated relevant issues. As such, an EyeDetect test determines a person’s behavior on an issue, not their guilt on one specific event. Responses to the relevant issue are compared by the algorithm to responses about a secondary issue.
A polygraph test that attempts to determine behavior on multiple, potentially unrelated issues, is less accurate because in that exploratory polygraph interview, the examinee may be questioned about separate issues for which s/he may be guilty on some issues and not on other issues. In those cases, there are too many potentially unrelated issues being compared.
EyeDetect is more accurate for screening (determining behavior) because of the use of comparison questions related to two similar, but likely unrelated issues.
While reading a computer-based series of questions, EyeDetect measures an examinee’s response accuracy, response time, pupil diameter, reading behavior and blink rates. The results are consistent with the cognitive workload hypothesis.Observations include:Guilty individuals — when compared to innocent individuals and while responding to simple test statements — make more errors, take longer to respond, make more fixations on the text, have longer reading times, and have longer rereading times.
Guilty individuals blink significantly less often as they process statements answered deceptively versus answered truthfully.
Guilty participants show greater increases in pupil diameter for statements answered deceptively than for statements answered truthfully.
Guilty participants respond faster, make fewer fixations, and spend less time reading and reading statements about the crime they committed than statements about another crime or neutral statements.
Simply put, an increase in the cognitive load is associated with recalling a task and used to distinguish between deceptive and non-¬deceptive responses. This is more pronounced when deceptive individuals respond to complex statements. It takes motivation and effort to deceive.
EyeDetect Security
Yes. If an organization wishes to protect the identity of an examinee for purposes of reporting testing results, during registration prior to taking a test, the Test Proctor would provide an identifying number rather than a person’s name. In addition, the Test Proctor can elect to not take a photo of the person being tested. This way, the organization will need to find test results for a person based on the assigned ID number–after the test is taken, saved, and scored.
To store and process user data collected during testing, Converus uses standalone (non-hosted) servers that are owned by Converus, and not owned by the data center. Access to these servers is controlled by a firewall and incoming web traffic is monitored for threats. All servers are housed in a private, locked rack in a SSAE 16/ISAE 3402 certified data center. Access to the data center floor is controlled by key card and biometric scanners and is monitored 24/7.
The Converus Dashboard web application is only accessible using two-factor authentication. All access to the dashboard is done through SSL. Only authorized users of an account with applicable rights can access the dashboard.
How secure is the data stored on the EyeDetect Station where tests are administered to the examinee?
Each EyeDetect station comes with a secure external hard drive manufactured by DataLocker. The DataLocker is 256-bit AES encrypted and FIPS 140-2 validated.
All test data is stored on the DataLocker and can only be accessed by users that enter the key to unlock the hard drive. Test data is encrypted using a unique key per customer before being transferred to our secure data center. Once the data is transferred to the data center it is deleted from the DataLocker.
EyeDetect & The Polygraph
Although the EPPA law specifically mentions polygraph, the law has been applied in practice more broadly to the use of any lie detection solution. At the heart of the law is the concept of protecting a person’s right to privacy in employment. Thus, any lie detector would be prohibited. However, lie detection solutions can be used today for screening or evaluating federal, state or local government employees. This includes law enforcement and national security agencies.
The absence of probable-lie comparison questions in an EyeDetect test eliminates concerns about any overlap that might occur between “relevant” and “comparison” questions (see “How is the polygraph typically used? What is the average accuracy for those uses?” above). The reactions to two sets of “relevant” questions are compared and each relevant issue serves as a control for the other issue.
An EyeDetect exam is almost completely automated and is administered by a computer. Therefore, the validity of the test does not depend on the interview skills of the examiner or their ability to properly interpret the physiological recordings. In contrast to a polygraph, the examinee for the EyeDetect exam does not attempt to appear truthful to an examiner.
EyeDetect tests can be developed for any language, allowing testing in the native tongue of an examinee. This eliminates the need for a translator and prevents language misunderstandings.
An EyeDetect test is less invasive than a polygraph test. During a polygraph, a variety of sensors are attached and some may cause discomfort. For EyeDetect, because a remote eye tracker records eye behavior during the test, there are no sensors attached to the examinee.
An EyeDetect tests takes about 40 minutes, in contrast to a polygraph examination, which may take 2-4 hours. An organization can conduct up to six EyeDetect tests in the time it takes to conduct one polygraph test.
Note: EyeDetect has a limitation that is not an issue for polygraph tests ― the examinee must have proficient reading skills.
When a deceptive person is devising a story to tell a lie, there’s an increase in cognitive load, or brain activity. If someone has the capacity to create a story without increasing how the brain reacts to this process, then that person will appear less deceptive. The pupils react involuntarily to an increase in cognitive load and for that reason EyeDetect has shown a superior level of accuracy compared to all other lie detection solutions on the market used to pre-screen job candidates and to conduct periodic evaluations of current employees.
If EyeDetect and polygraph are used together, the probability of a false negative error is greatly diminished. Note: False negative means a deceptive person is categorized as truthful.
EyeDetect primarily measures cognitive processes and the polygraph primarily measures emotional responses. As such, the two tests would be relatively independent and would provide complementary information about a person’s deceptive status. In the case of guilty individuals, both tests aim to keep false negatives at a minimum. When used in combination, EyeDetect and the polygraph can achieve accuracy rates of roughly 96.6 percent.
To determine that figure, take the product of the probabilities of false negative errors for the two tests. For example, if the probability of a false negative error is 17% for EyeDetect and 20% for polygraph, the joint probability that a deceptive applicant would pass both tests would be .17 X .20 = .034 or 3.4%. That’s 96.6% accurate.
The polygraph exam typically has two main uses: for incident-specific questioning, like the police use in investigations; and for screenings, such as the ones employed by government agencies.
1) In criminal investigations in the U.S., the polygraph is used to determine if statements by suspects concerning a specific crime are truthful or deceptive. The polygraph test in this case is specific to the incident in question. There are two types of questions asked. First, examiners will ask “relevant” questions pertaining to the specific incident under investigation, such as, “Did you rob the Quik Mart on June 14, 2014?” Examiners also use “comparative” questions (called probable lie questions), which are used to establish a reaction to intentionally vague or difficult questions, or are questions that are impossible to answer truthfully with an unqualified negative answer. For example, one might ask, “Between the ages of 18 and 28, did you ever lie to someone in authority?”
At the conclusion of the test, the polygraph examiner compares the physiological reactions to the various question types. A subject who reacts more strongly to relevant questions is considered deceptive. The accuracy of this type of test is influenced by the skills of the examiner. Under optimal conditions, the accuracy of probable-lie tests in incident-specific criminal investigations is approximately 90 percent. (Source: American Polygraph Association Ad Hoc Committee on Polygraph Techniques, 2011)
2) Various U.S. government agencies use polygraph test to screen job applicants, employees, sex offenders and parolees. In contrast to incident-specific criminal investigations, relevant issues are more general, such as: “In the past 90 days, have you used any illegal drugs?” The generality of relevant questions in screening examinations is considered desirable because the test covers a wide range of illicit behaviors of concern to the agency. However, the generality of the relevant questions may introduce ambiguity in the mind of the examinee about their guilt (“I haven’t used illegal drugs in past 90 days, but I used them six months ago, and I know that was wrong”). The generality of relevant questions also increases their similarity to comparison questions, which are intentionally vague and broad in scope.
It is expected that reactions to comparison and relevant questions are more similar in magnitude and less diagnostic in screening tests than specific-incident tests, increasing the risk of false positive and false negative decision errors. Under these circumstances, the accuracy of polygraph falls in range of 65 to 85 percent.
EyeDetect can be used to replace in some circumstances and/or can complement the polygraph. In reality, EyeDetect is the perfect add-on service for a polygraph company. Studies show that the polygraph can be very accurate for event-specific questioning (a specific line of questioning). However, studies also show that EyeDetect has shown a superior level of accuracy compared to polygraph when used for employee pre-screening and periodic evaluations.
Polygraph machines sense changes in glucocorticoid (hormone) levels, which initiate chemical reactions such as changes in respiration, heart rate and skin conductance (a moisture buildup under skin). Thus, a polygraph measures physiological changes when under the stress of questioning. That questioning is intended to determine whether a person is lying or being truthful. EyeDetect measures changes in pupil size and eye movements that reflect the changes in their brain activity while a person reads and responds to a questionnaire.
For the polygraph, the primary testing premise is that a deceptive person will show stronger emotional reactions to questions about topics for which they are deceptive. For EyeDetect, the primary testing premise is that a deceptive person will show an increase in cognitive load when questioned about topics for which they are deceptive.
Lying
Other Deception Methods
This type of test is based on the premise that muscles in the voice box tighten or loosen, which changes the sound of the voice. The small, involuntary frequency modulations in the voice when under stress (i.e., lying) are measured. Studies show that this type of tests operates at about a chance level or slightly higher (50-65% accurate). That means it would be similar to a flip of a coin. Source: Journal of Forensic Sciences, 53(1), 183-193, Hollien and Harnsberger (2008)
Aptitude tests determine if a person is apt to do the job. In other words, can the candidate perform a job function? Predicting if someone will do the job is a challenge. Predicting if someone will lie, cheat or steal is also a challenge. The premise behind EyeDetect is that is indicates recent, past behavior. The employer can determine how to use that information to take action with an employee.
Personality tests try to determine if there is a job fit from a behavioral perspective. The challenge with a personality test is that there are no right or wrong answers. The test will assess a person’s dominance, altruism, neuroticism, egotism, psychopathy, introversion, etc. or other items. Some personality tests try to assess a person’s willingness to pretend to be good. Employment personality tests try to determine a person’s interaction style and behavioral tendencies. They attempt to assess aspects of personality that remain relatively stable throughout a lifetime. They also attempt to predict future behavior.
Integrity tests assess the likelihood an employee will engage in dishonest behavior. This is an attempt to predict future behavior. An “overt” integrity test discusses past criminal behavior and attitudes about honesty, drug use, theft and counterproductive behavior. If a person is completely honest, this test can be very valuable. Deceptive individuals may not be forthright in their responses.
One notable study entitled “The Use of Integrity Tests for Pre-employment Screening” discusses the validity of integrity tests. The study was published by the Office of Technology Assessment (OTA) of the U.S. Congress.
Some points of interest follow:
STUDIES ON WORKPLACE THEFT
The results from 5 studies on workplace theft showed that integrity tests misclassified dishonest persons as honest between slightly less than 1% to 6% of the time. The tests also misclassified honest people as dishonest from 73 to 97% of the time. In summary, workers were misclassified between 18% to 64% of the time. (pp 10-11)
STUDIES ON COUNTERPRODUCTIVE BEHAVIOR AT WORK
The results from 3 studies on counterproductive behavior showed that integrity tests misclassified counterproductive persons as productive between 18 to 29% of the time. Two of these studies also misclassified productive people as counterproductive from 22 to 29% of the time. (p 11)
HONEST PEOPLE FAIL
Other research showed that between 30% and 60% of all applicants will “fail” an integrity test (categorized as dishonest). There will be an important percentage of honest people that are turned away. (p. 12)
Studies say that humans have an accuracy rate of 54% for catching liars. (Source: Bond & DePaulo, 2006) If that’s true, you’ll get the same results from chance, or in other words, from the flip of a coin.
Behaviors such as gaze aversion, touching the body or face, or covering the eyes or mouth while speaking have not been found to be reliable indicators for deception.
Interested?