 Hello and welcome to this presentation. Today we're going to talk about eye-detect test structure and test format. My name is Russ Warner, I'm with Converis. If you have questions after this presentation, you can email me at rwarner at converis.com. In this photo you see Dr. John Kircher and on the right you see Douglas Hacker and in the middle Don Crapple. As these men were traveling to Mount Rainier from Utah to go on a hiking trip of sorts, they were discussing the topic of the effect of cognitive load on eye behavior and they determined that they were going to return to the lab and do this research based on this basic premise that as a person thinks more, their eyes respond. And in this case, if you could put the person in front of a computerized system that would ask questions about target behaviors and then at the moment that they were responding about those behaviors, if you could detect changes in cognitive load which are manifested involuntarily in eye behavior, then you might have something. And that's basically what happened. Eye-detect was developed over a period of 14 years and during that 14 years the process was devised and created which essentially is an automated test delivered to the examinee by a computer. The test is administered by a test proctor which is a person who simply adjusts the monitor, sits the examinee in front of it, and begins, launches the test process on screen. As the examinee takes the test, the test proctor merely observes their behavior and makes sure that there are no questions. At the end of the test, the examinee is escorted out of the room, the test proctor then uploads the data which are basically eye measurements and true and false responses gathered during the test and those are uploaded to a cloud server where an algorithm scores the test. So in essence, there is very little contact between the examinee, well, there is zero contact between the examinee and the person scoring the test because it's done by algorithm and in terms of having contact between the examinee and the person administering the test, there's very little contact. So the test is predetermined. The structure has been created, the test types and test questions have been predetermined. The basic structure of the test is called the relevant comparison test, which I'm going to discuss in a moment, but this whole process has been automated. It removes human error because it is automated. It is more standardized and delivered the same way to every examinee in the same fashion, in the same structure, in the same process. By doing that, you add objectivity. You reduce subjectivity because the test is administered the same way, in the same fashion, to everyone. That increases precision and accuracy and because there's no contact with the person scoring the test and the examinee, bias is reduced. The standard process or testing process for an eye-detect test is that the examinee sits down at the computer, the test begins, pre-test instructions are given using mind maps, which we'll talk about. Instructions are essentially read to the examinee, which gives them the overview of what the test, topic, or test content is about. The examinee then takes a practice test, which allows them to become familiar with the testing process, with how to respond to the questions, how to push the buttons on the keyboard, etc. So they don't have to learn the process while being tested. For basically topics that have consequence in many cases, after the pre-test instructions and the practice test, the standard credibility test then is carried out. I mentioned that mind maps are used. This is a concept that was created in the days of polygraph and are still used today. It is a graphical representation of the topics to be discussed. By seeing and hearing the instructions on screen, the examinee will retain more and be more familiar with the topics and there should be more jeopardy because there's more understanding. Every eye-detect test has a relevant or a primary issue and a secondary or comparison issue. We refer to the primary issue as the R1 issue and the secondary or comparison issue as the R2 issue. A primary issue could include drug use, for example. That's a single issue, primary issue or R1. A test could also include as the R1 issue three topics, for example. In a typical law enforcement pre-employment test, the test would discuss drug use, involvement in serious crimes, and possibly falsifying the police department application. That would be three topics combined into one for a multi-issue test. So there are a single issue test or a multi-issue test and those issues are the primary issue or the issue of most concern. There's always a secondary issue, though, in every eye-detect test, which is a comparison issue. It essentially helps establish a baseline in terms of the reaction of the examinee to a topic or a target behavior of concern, but really the R2 issue is a serious crime with a very low base rate of occurrence, which should have face validity to the examinee. In other words, they must believe that the organization cares about the issue, severe nature and consequence, called the probable truth issue by polygraph examiners. This issue should invoke a response from the examinee if they participate in the behavior, but because we choose a very low base rate type of crime, very few examinees will have actually participated in the behavior. As a result, when you establish a baseline reaction for the examinee against a serious crime or serious consequences and see no reaction, then when you're asking the examinee about the primary issue and there are reactions, you get this differential type of reaction and that helps eye-detect determine deception in examinees. In addition to the R1 and R2 issues asked in a test, also we use neutral questions, which are trivia type questions, world knowledge questions, which basically change the topic for the examinee. It allows the examinee to stop thinking about the primary issues or target issues or target behaviors for a moment while they're responding to other question types. And we also use alpha arithmetic questions, which are simple math questions, which also change the topic for the examinee and also show a different level of cognitive load reaction to these simple math problems. As I mentioned, you can have single-issue tests, for example, testing about drug use, for example, or you can have multi-issue tests. In the case of a law enforcement pre-employment test, we could ask about drug use, serious crime involvement, and falsifying the police department application. Every one of these tests or every test that's created for eye-detect can be customized for the end user or the organization. The test instructions could mention the organization. It could talk about the job description of the person taking the test and it will then reveal the target behaviors of concern and then the test is administered. When we carry out a multi-issue test, it's very important to use audio-visual instructions with those mind maps that we mentioned earlier to get the concepts more solidified into the mind of the examinee. Eye-detect is also used on occasion for screening groups, generally for investigation purposes. If there's been a crime in an organization, you could use eye-detect to screen all of those employees, perhaps, who had access to those items stolen in this case. But we wouldn't typically use eye-detect in a specific investigation where you would ask about a specific crime on a specific date and time, for example. In the case of a screening test for a group under investigation, the test might more generally ask about stealing from the employer over the last three months. So eye-detect is used for pre-employment screening, for periodic employee reviews, and for screening groups, generally speaking, during an investigation. Topics for eye-detect tests can be, as I mentioned, single issue, stealing, or there can be multiple issues, which would combine, as you see on screen, any one of these target behaviors of interest. You could ask about drug use and crimes and counterfeiting, for example. Combining the test topics is easy. We simply pick the test topics to be combined and then craft or design a test around those topics, choosing an R2 issue as well. In the case of these, these are some examples that you see on screen. Eye-detect is used around the world, and in some countries, there are no limitations to the types of lie detection tests that can be administered to private company employees. In the United States, however, because of the Employee Polygraph Protection Act, lie detection tests cannot be administered to private company employees, but only to federal, state, local, or municipal employees. In this case of the examples you see on screen, you can see that we have created eye-detect tests for a variety of job descriptions and for a variety of target behaviors. We have created law enforcement pre-employment tests, both inside and outside the US. We've created intelligence agency tests to discuss with the examinee their acts of espionage, for example, or sabotage. For a security guard or armored car company, we've created tests that ask the examinee about drug use, involvement in serious crimes, and maybe counterfeiting, because if a person is going to have contact with cash, you want to know if they have laundered money in the past. And as you see on screen, there are many other types of or examples of different tests. So basically, eye detect can be very customized. It can be tailored to the job description and to the organization. Some examples might include, and this is an example of the test topic combinations that are available. You could create a multi-issue test that talks about drug use, crimes and ties to criminals, which would be organized crime or cartels, and then ask as a comparison issue about drug trafficking, or as a different comparison issue, ask about violent crimes. So you can take any type, any test topics, combine them for the relevant, the relevant issue and then choose a serious crime with a low base rate of occurrence and use that as the R2 or comparison issue. In the test, pretest instructions are given to the examinee. There's an explanation of the primary issue, the R1 issue, and an explanation of the comparison issue or the R2 issue. Examples are given of each type to clarify and make certain that the examinee is on the same page, so to speak, as the organization. If there are multiple issues to be discussed, they're each defined. And if the examinee participates, we would say, in any of those target behaviors, then he or she would be disqualified from the position for which they're interviewing. So a multi-issue test essentially combines all three issues into one question, which says that if you're involved in any of these target behaviors, you will be disqualified from the process. As an example, if we're going to use serious crimes as one of the R1 issues, we would explain to the examinee during the pretest instructions that today we are concerned about their involvement in serious crimes. A serious crime is a crime against people or property. We want to know about crimes whether they were caught or not caught. In other words, detected or not detected. These are the types of crimes that if you're caught, they could result in your arrest, your conviction in fine or going to prison. And essentially, you're describing for the examinee very serious crimes. But in order to make it even more clear, you give examples and each organization has different examples. But for example, you could say, examples of these types of crimes include theft, assault, kidnapping, fraud, sexual assault, drug trafficking, and so on. And if you participate in any of these crimes, you say to the examinee, you'll be disqualified from this job, process this job or hiring process. If the issue was drug use in the pretest instructions, you would explain to the examinee that today we're concerned about their involvement in drugs. And specifically, we care about drug use in the past 24 months, not prior to that, but basically in the last two years, examples of drugs that we are concerned about include marijuana heroin cocaine, methamphetamines, crack hallucinogens, and you can list any other types of drugs that you'd like, or and even use the catch all phrase that you would also that the company is also concerned about any other illicit drug use. And if the examinee is involved in any of these drug in any of this drug use, in the past 24 months, they will be disqualified from the hiring process. In the case of the questions that are asked during the test, we call them statements, they're true and false statements. The examinee must listen, sorry, it must read the questions on screen, consider the response and then respond true or false. All the questions are written with the innocent examinee in mind. In other words, as you see on screen here, these are four examples of our one statements about serious crimes. The first one says, I was involved in illegal activity at a previous company. If you're innocent, your answer to that question would be false. I was not involved. Question number two, I did not participate in illicit behavior at a previous company or previous employer. If you're an innocent examinee, the correct response to that question would be true. And each one of these questions is written in a slightly different way. In the case of number three, in the past illegal activity was never carried out by me at work is the as an example of a passive voice question. Question number two has the word not in it. That would mean that the question has a negation in it. Question number one is written in active voice, first person. And so each of these test questions are essentially written in a variety of different formats using active voice, passive voice, sometimes using negations. And if the answer to the question is false, and if the question included the word not or never in it, that would be a double negative. Now, although I detect doesn't rely on double negatives, it's sometimes interesting to include at least one double negative question, because you want the examinee to read the read the question, consider the response and then respond by check, pushing on one of the keys on the keyboard. By doing so, you get them to consider thoughtfully, each and every answer to each and every question. For this reason, all of the test questions are not identical. If they were, the examinee would begin to habituate. And they would think less and less and less about the question as they read them. Here are some examples of our two statements, or statements for the R2 issue that have to do with counterfeiting. I have not been involved with counterfeiting money. The answer to that question would be true if you're an innocent participant. That question includes the word not, therefore, it's a negation. Crimes with counterfeit money are part of my past, or I have not used counterfeited money to commit a crime. Some of my past transactions involve counterfeit money. This is only for examples. In the case of the R1 questions and the case of the R2 questions, there are 16 different question formats. And the examinee will see the 16 different question formats, half of which are true responses, or eight of the questions are true responses, and eight of the questions are false responses. By having this variety of question types, it elicits a thoughtful response from the examinee. Some of the neutral statements would include examples such as true and false. I am less than 95 years old. True and false. A triangle is a geometric figure with four sides, and so on. These questions are typically, well these questions are neutral statements. They're designed to get the examinee to stop thinking about the topics at hand, which would be crimes typically, and to think about something different for a moment. These are simple examples of the neutral questions or neutral statements. Finally, we have samples of the alpha arithmetic statements, which are basic math problems. B plus one letter equals the letter C. True or false? The answer to that would be true. The letter B plus one letter of the alphabet would be the letter C. These are simple examples, and if you decide that the examinee pool that you're testing would find these examples to be too difficult, you can, we can modify the test to include even more simple math questions than the ones that you see on screen. So when you take the test, the examinee sits down and they hear the pre-test instructions, which we also call a preamble. That process takes a little bit less than five minutes. After that, the examinee will take the practice test to learn how to take the test. Then after that, the lie detection or truth verification test will begin, and that lasts about 25 minutes. All in all, the test takes about 30 minutes. As I mentioned, the true and false statements could be written in simple, in active voice or in passive voice, simple statements, sometimes using negation. And an algorithm is used to measure the exact length of the statement, the number of characters and spaces. And then it combines in the algorithm an indicator or a consideration of the complexity of the statement. If the statement has a negation, the examinee will be given slightly longer to consider the response. If the example, sorry, if the statement is written in active voice and it's very straightforward, the examinee will have slightly less time to respond. So each of the question types is reviewed, a time frame is calculated, and the examinee will be given that amount of time in the test to read and then respond. For this reason, we know exactly how long the eye-detect test will last, because every single question has a maximum amount of time in which the examinee can respond. And that maximum amount of time to read and respond is generally up to 10 seconds. Most examinees don't need that much time. The average would be about 3 to 4 seconds. So the test duration is about 2 minutes. In the test, we have two practice sessions, as I mentioned, to get the examinee familiar with the process. Then during the test there'll be five sessions of questions which last about three to four minutes each. And then in each of those five sessions, the examinee will see R1 and R2 and neutral questions all in a random order. In between those five sessions of R1, R2, and neutral statements, the examinee will have four sessions of just alpha arithmetic statements. All in all, there are nine sessions. Each with a rest of 15 minutes or break of 15 seconds between each session. That way the examinee won't fatigue and the topics change significantly enough to keep them interested perhaps or involved in the test process. And after each one of those sessions, there's a telemetry step, which basically is a phase in which for those right after the 15 seconds, the screen will change. The eye tracker will then make sure that the examinee's eyes are still in the proper position and that data is still being gathered effectively because if an examinee begins to implement a countermeasure of squinting or looking away, eye detect can determine that they're doing this. But essentially the test takes 30 minutes, nine sessions of questions, a break in between each session and two practice sessions. All in all, there are over 300 statements or questions. There are 20 practice questions. There are 16 relevant issue questions or statements, 16 secondary or comparison issue questions or statements, 18 neutral statements, and 48 alpha arithmetic statements. And because there are five sessions of R1, R2 and neutral statements, the examinee will get five times 16 or 80 chances to deny their participation in the target behavior or 80 chances to indicate that they've not been involved in the target behavior if they haven't been. Possible applications for eye detect include creating, as I mentioned, pre-employment screening test, periodic test for employees, you could include immigration screening to ask immigrants for their participation in perhaps terrorist activities, sex offenders to ask a sex offender about their undisclosed victims, their involvement in the crime at hand, in parole violations for example. And you can use eye detect alone or in series or in combination with with polygraph because when you combine two independent tests that measure two different sets of physiological reactions, you can expect to get up to 99% outcome confidence in the results if the examinee passed both tests or failed both tests. So in combination with polygraph, eye detect can be used to be a very significant tool in the process of credibility assessment. If you have any questions, please let me know. I do appreciate your time. Thank you very much.