The Interview Challenge.
On a Wednesday morning, an FBI special agent spends approximately 1 hour interviewing a local bank's loan officer, Mike Simmen, about a theft from the bank's automated teller machine (ATM), which occurred the previous Saturday. After asking numerous questions, the agent believes that Mr. Simmen knows more about the theft than his answers convey. The agent decides to try a more direct approach. "Tell me what you think happened on Saturday," the agent asks. Mr. Simmen replies, "Well, I think somebody took the money." Next, the agent says, "Tell me your side of what happened Saturday." Mr. Simmen retorts, "What do you mean, my side? Do you think I took the money?" The agent reassures Mike, "Everyone remembers things differently, so that's why we need to get everyone's perspective." Mr. Simmen answers, "I don't know that I have one." Then, the agent pointedly remarks, "If you did it, we'll find out." At this, Mr. Simmen indignantly announces, "I'm not going to answer any more questions. This interview is over.
Realizing that this approach did not succeed, the agent decides to interview Mr. Simmen again, using a different approach. How could this happen? Thanks to modern technology, law enforcement officers now can enhance their most fundamental and important skill, interviewing individuals, via an interactive computer program that so closely imitates real life officers may find it difficult to tell the difference.
DEVELOPING THE PROGRAM
In 1996, instructors who teach interviewing and interrogation at the FBI Academy met with members of the Johns Hopkins University's Applied Physics Laboratory (APL) to determine if they could create a computer program that would realistically simulate a human personality.  At the time, these instructors were training record numbers of new agent and FBI National Academy students. They needed a system that would allow their students to practice interviewing skills and receive feedback yet did not require instructor-monitored practice sessions. Primarily, the instructors wanted a program that would augment and support the training that FBI students receive during their interviewing classes but that also could act as a standalone practice drill to enhance the interviewing skills of veteran officers.
The APL faced a challenging design proposal. First, the FBI instructors wanted an interactive, self-paced computer program user-friendly enough to allow those with minimal computer expertise to use it with little or no training and engaging enough to make students want to use it on their own time. Second, the computer-simulated interviewee needed to display multiple dispositions interview after interview to emulate the many different types of individuals that investigators encounter. Finally, the program could not allow users to "beat the system." FBI instructors did not want clever students devising one set of questions that they could ask in the same order during each interview to produce a high score. Human interviewees prove more complex than that.
The APL accepted the challenge and delivered the completed software to the FBI in May 1998. "Mike Simmen" (i.e., simulated man) was born. Mike may appear talkative and eager to help investigators in one interview but seem busy and defensive in the next. He may portray an innocent employee in some interviews but, in others, is guilty of stealing money from an ATM. As with many humans, even when he is not the perpetrator, Mike may lie to hide other information that he does not want the user to know. These changes occur because Mike "remembers" the nature of the user's questions and statements and responds based on typical behavior patterns related to his guilt or innocence and the content of the interview.
Moreover, the APL designed Mike's "brain" with both logical and emotional components. The logical component tracks the responses and keeps them reasonable and consistent. It selects one of a series of likely responses to the current questions and circumstances, which affect Mike's actual status (i.e., guilt or innocence) and emotional state. At the same time, the emotional component critically impacts Mike's response selection. While the user's questions primarily determine Mike's emotional state, the computer randomly selects the fluctuations of Mike's emotional state or "mood," causing his answers to change each time the user conducts an interview. For example, depending on Mike's mood, he may forgive a poorly worded question or become upset and uncooperative. The user never knows how Mike will respond from one interview to the next.
USING THE PROGRAM
The program includes an online tutorial to help users learn how to use the program and to understand the scoring system. Also, an online manual of tips and guidelines helps those who want to enhance their interviewing skills. For example, the manual reminds users that an interview is a conversation with a purpose or goal, not just a series of questions. To this end, the manual stresses that users must learn to evaluate the truthfulness of the information they obtain by "reading" both the verbal and nonverbal indicators of the individuals they interview. As experienced interviewers know, they must develop rapport with their interviewees and establish a baseline of what constitutes an individual's normal behavior. Without determining an individual's typical reactions, interviewers cannot identify deviations from them. 
Conducting the Interview
After reviewing the online case study to obtain background information about the crime, users start the interview by choosing 1 of 14 different categories, covering such items as Mike's personal habits, work relationships, or possible involvement in the crime. Users then conduct the interview by selecting from an extensive scripted list of questions. As Mike responds to these inquiries, additional follow-up questions appear. Users choose the questions that they feel are most appropriate. Simultaneously, the program eliminates those questions that users have asked or those that have lost their relevance.
As the questions and Mike's responses appear in a portion of the computer screen, users see a full-body view of Mike seated in front of them and a close-up of his face in another part of the screen. At the same time, users hear their questions followed by Mike's responses. While Mike's "brain" determines his behavior and responses, an actor presents the visual and audible responses in the video sequences. This simultaneous visual and aural presentation realistically simulates a lifelike interview. Unlike an actual interview, however, the program stores the sequence of questions and responses so that users can replay and reexamine the entire interview at any time until it ends.
As the interview progresses, skilled interviewers recognize that some of Mike's verbal responses and nonverbal body movements readily indicate guilt or innocence and represent clues to his level of truthfulness. After hearing and observing Mike's responses, users can plan a line of questioning to help them judge these verbal and nonverbal clues as truthful or deceptive. For example, users must decide whether Mike's tone of voice, body posture, and other nonverbal actions, such as scratching the back of his neck or avoiding eye contact, demonstrate deception. To this end, if users have not determined Mike's normal behavioral reactions correctly, they may believe his lies or suspect him when he is innocent. Also, similar to computer and video games, the users gain extra points for successfully identifying these truthful or deceptive clues but have points taken away for incorrectly judging Mike's responses. This encourages users to remember important interviewing skills, such as developing rapport and establish ing a baseline of normal behavior patterns.
Concluding the Interview
Unless Mike refuses to answer any more questions (as he did in the beginning scenario), users decide when to end the interview. At that time, they must determine whether Mike was truthful or deceptive. Users must base their decisions about Mike's truthfulness on the complex combination of information they received from Mike in response to the questions they asked throughout the interview. This often proves challenging. For example, if users appear too abrasive in their quest to solve the crime, Mike may end the interview when he becomes frustrated. Yet, if users fail to probe sufficiently, they may find it difficult to determine Mike's truthfulness.
Once users decide whether Mike appeared deceptive or truthful, they must enter their decisions. Then, the computer program gives them information about their interviewing efforts.
Scoring the Interview
The program computes a total numerical score based on the accuracy of the users' decisions concerning Mike's truthfulness, the rapport and investigative values of the questions they selected, the number of questions asked, and the number of clues detected correctly. The program has four levels of difficulty--beginner, intermediate, advanced, or professional--and provides fewer clues at the more challenging stages.
To illustrate one of these rating factors, one objective of this program involves users' developing rapport with Mike and, thus, more likely to provide reliable information. Accordingly, the program rates each of the users' questions or statements on how it contributes to this rapport. Positive rapport-building questions generate positive rapport ratings, whereas questions that offend Mike contribute to negative rapport-building scores. However, some of the questions that make Mike feel uncomfortable may help the user determine Mike's deceptiveness and provide important investigative information. Therefore, both rapport and investigative ratings contribute to users' overall evaluations.
Similar to video and computer games, the program compares users' scores and lists the top five  along with the date they occurred. Users may print a copy of their scores, save the interview for later replay and further examination, start a new interview, or exit the program.
MEASURING THE PROGRAM'S SUCCESS
Since October 1, 1998, all FBI new agent trainees have used this interviewing software to augment their classroom instruction.  Although not enough time has elapsed to calculate any notable benefits, FBI Academy instructors have observed improvement in the interviewing skills of those who have used the program. For example, most trainees who have practiced with it conduct longer, more in-depth interviews during classroom role-plays. Many spend additional time building rapport by developing general conversations about nonthreatening topics (e.g., family, education, the weather, or sports); employing more thoughtful questioning strategies, such as asking open-ended questions (e.g., What happened? Can you tell me what you have heard?); and encouraging conversation by asking follow-up questions (e.g., And then what happened? What else did you hear?).
Additionally, FBI instructors require that new agent trainees use the program for two homework assignments. At the beginning of training and then near the end, the trainees must interview Mike, as often as they wish within the assignment's time limit, and submit their best score of each of these interviews to their instructor. Finally, most of the instructors agree that while nothing replaces interviewing a real subject, this program allows students to make mistakes and, more important, learn from their mistakes in a supportive instructional environment.
EXPLORING OTHER APPLICATIONS
Such an absorbing and interactive training tool has a broad range of potential applications. The sophistication of the simulation and programming of the software makes it a viable candidate for many kinds of training that require understanding the psychology of human interaction. For example, any investigative agency could use the basic approach developed for the FBI to detect deception in a variety of situations, including employment interviews, security investigations, and insurance claims. However, the program also has applications in many other areas. Schools, businesses, and community service organizations could adapt it to teach individuals how to respond in specific situations, such as showing young people how to resist peer pressure and other unhealthy influences or informing supervisors how to interact more effectively with their employees. 
OBTAINING THE PROGRAM
Early in 1999, the FBI Academy conducted interview training for veteran FBI agents from most of the agency's 56 field offices. Attendees used this software during these seminars and took copies back to their offices for other agents to examine. Afterward, the FBI provided numerous copies of this software to each of its field offices for distribution to state and local law enforcement agencies. Departments can obtain the software at no cost by contacting the nearest FBI field office.
Solid interviewing skills stand as the cornerstone in law enforcement's arsenal of crime-fighting weapons. Officers need to sharpen these abilities as surely as they must hone their expertise with firearms. Accordingly, advanced technology has provided an innovative and engaging training tool that law enforcement administrators may want to explore to help their officers enhance their interviewing skills.
FBI Academy interviewing instructors worked with the Johns Hopkins University's Applied Physics Laboratory to provide law enforcement professionals with realistic interviewing practice via a self-paced, multimedia computer program. The resulting user-friendly software gives students and veteran officers alike experience in asking appropriate questions and distinguishing between deceptive and truthful responses. It also provides a critique and numerical score for users to compare their level of improvement. This creative response to a critical need in the law enforcement community represents how modern technology can help officers in their daily struggle to successfully solve crimes and safeguard the communities they are sworn to protect.
Special Agent Einspahr is an instructor in the Law Enforcement Communication Unit at the FBI Academy.
(1.) Prior to this meeting, the APL had demonstrated an extremely lifelike, interactive training computer program to FBI Academy personnel. Following an assessment to determine the area of instruction that would benefit most from the APL's assistance, FBI Academy managers chose interviewing because of its fundamental importance to police work.
(2.) For additional information on conducting interviews, see, for example, David Vessel, "Conducting Successful Interrogations," FBI Law Enforcement Bulletin, October 1998, 1-6; and Michael R. Napier and Susan H. Adams, "Magic Words to Obtain Confessions," FBI Law Enforcement Bulletin, October 1998, 11-15.
(3.) Many users want to know the highest possible score. However, the extensive number of human variables makes it impossible to determine or achieve an absolute top score.
(4.) Because the software only recently became available in sufficient quantities for distribution to state and local law enforcement agencies, FBI National Academy students began using the program in October 1999.
(5.) For complete details about different software applications, contact Dale E. Olsen, program development director for law enforcement programs, Applied Physics Laboratory, Johns Hopkins University, Laurel, MD; telephone: 443-778-6114 or 240-228-6114; e-mail: firstname.lastname@example.org.
Minimum System Requirements
The program will operate on a standard, up-to-date desktop or laptop computer that meets the following requirements.
Graphic card support and screen setting of 1024x768 pixels or more with 16 bit color, 32 megabytes RAM, Pentium Processor 180 MHZ or better
* Hard Drive Option: 600 megabytes of free disk space
* CD-ROM Option: 40 megabytes of free disk space, 12x CD-ROM
* windows 95 Option: Microsoft DirectX 5.2, Microsoft Active Movie 2.0, and Microsoft Internet Explorer 4.0
* Windows NT Option: Version 4.0 with service pack 3 (contains DirectX 3.0), Microsoft Active Movie 2.0, and Microsoft Internet Explorer 4.0
|Printer friendly Cite/link Email Feedback|
|Publication:||The FBI Law Enforcement Bulletin|
|Date:||Apr 1, 2000|
|Previous Article:||Successful Adjudication Partnerships.|
|Next Article:||Criminal Investigation Handbook: Strategy, Law, and Science.|