The military has moved from a traditional view of intelligence gathering as the primary responsibility of intelligence specialists to a perspective in which “every soldier is a collector” (Jackson, 2007), where every soldier participates in intelligence gathering as the eyes and ears of the force. For example, Hazlett (2006) has noted that the human collector remains the most deployable and adaptable tool that can be put in the field. Accordingly, good intelligence is built up from the gathering of low-grade or field intelligence that takes place at checkpoints and in the street-level encounters between military personnel and civilians.

There are two consequences of this “every soldier a collector” perspective. First, it emphasizes the gathering of intelligence in situ or in place in the field settings in which military operations occur. In a typical scenario, military personnel may pull a vehicle over at a checkpoint and engage the passengers in a short conversation as to where they have been or what activities they have been involved in, or a soldier may have the opportunity to question family members or persons of interest on a street corner. Second, these brief investigative interviews take place in a social context, which is likely to involve two or more persons of interest. That is, the field interview may take place on a one-on-one basis but is just as likely to include the opportunity to question two witnesses or accomplices or co-conspirators together.

We are reminded of the haunting video image of the two Boston marathon attackers walking together in the crowd prior to the bombing. If an official had reason to engage them at that point, would there be indicators in their responses to questioning that would have lead the interviewer to perhaps detain them for further questioning? In other words, are there unique cues to deception that may be observed between co-conspirators or accomplices? Are there social indicators of deception?

Why Is This Question Important?

A primary reason this topic is important is that it relates to the practical requirement for understanding deception at a social level. Many opportunities for intelligence gathering in the field occur at the social or group level. In describing cordon and search operations in Iraq that uncovered an improvised explosive device manufacturing cell, a platoon sergeant stated “I could tell by the two guys standing outside. They didn't seem right” (Miller, 2005, para. 8). The important question from a scientific standpoint is as follows: Why didn't they seem right? What did they say or do that seemed deceptive? In other words, if you question these two persons jointly regarding their activities, are there characteristics of speech or behavior that are exhibited between the two suspects that indicate deception?

A second reason this topic is important is that we have almost no existing scientific data available to answer this question. Regrettably, research has failed to keep up with changes in intelligence gathering practice. A considerable amount of research over the years has examined the individual act of deception, lying, or giving false information. In a typical study, an individual commits a transgression and then lies to an interviewer whose task is to discern credibility. One significant limitation of this research is that it examines deception as an exclusively individual-level phenomenon. Most of this research has examined the behavior of either a single deceiver (i.e., research to determine the cues exhibited by individuals during deception) or a single observer (i.e., research to assess the cues used by observers to detect deception).

Table 1 illustrates the state of the research literature. Almost all research on deception is represented in Cell A: studies in which an individual commits a transgression and is interviewed individually to determine truth or deception. Cell B is not relevant. In Cell C, there are a handful of studies that have been conducted in which two or more persons jointly participate in a transgression, yet they are isolated and interviewed individually. In Cell D, in which two or more persons have jointly participated in a transgression (which we term “conspiracy”) and in which information is obtained by interviewing these persons jointly, research is almost nonexistent.

There are several reasons why this is the case. First, much research on deception has taken place in academic settings in the experimental laboratory, conducted by researchers primarily interested in individual emotions, such as guilt or shame. Second, there is a longstanding tradition within law enforcement to isolate potential suspects as soon as possible prior to interviewing or interrogation.

Isolation is carried out to remove the subject from familiar surroundings and people, heighten the stress of interrogation, and increase the subject's anxiety and incentive to confess. Yet, some have argued that there is an important difference between the law enforcement interrogation and intelligence interviews: The purpose of a law enforcement interrogation is to obtain a confession from a suspect, whereas the purpose of the interview in an intelligence-gathering context is to gather accurate, useful information from a source.

The result is that we know very little about social indicators of deception—unique cues to deception that may occur between co-conspirators or accomplices. The goal of our initial research was to extend the study of deception beyond the analysis of individual deception to situations in which two or more people may be involved in a transgression and in which information may be obtained by interviewing these persons jointly. The theoretical contribution of this research is that it extends existing research that only looks at individual deception by considering deception in a social context.

What did we hope to gain from this perspective? First, although DePaulo et al. (2003) noted that “the looks and sounds of deceit are faint” (p. 104), the looks and sounds of deceit may be stronger in a situation in which one suspected conspirator is in the presence of another. In other words, it is likely that the signs of deception or intent that are evident on an individual basis may be more evident when conspirators are in the presence of one another. There is some theoretical basis for predicting that the presence of a co-conspirator may lead to an increase or escalation in signs of arousal/affect (Mullen, Bryant, & Driskell, 1997).

Second, there may be different types of verbal or nonverbal cues that distinguish between individual deception and conspiracy. For example, gaze may not be a reliable indicator of individual deception when operationalized as the extent of gaze between a suspect and interviewer (see DePaulo et al., 2003). Yet, gaze may be a reliable cue to conspiracy when operationalized as the extent of gaze between two co-conspirators under questioning.

Third, there may be “looks and sounds” of deceit at a group level, such as cues stemming from interaction between co-conspirators, that may not be apparent when these persons are interviewed or observed individually. For example, interactive behaviors—such as acknowledgments, back-channel responses, or corrections or repairs of communication—may serve as cues to deception during interaction between suspects, and these potential cues are simply not observable at an individual level of analysis.

Study 1

To test this approach, we conducted an initial study to examine social indicators of deception (J. E. Driskell, Salas, & Driskell, 2012). When people lie, they often concoct a story about an event that did not take place. For example, if you ask two suspects whether they had been in a restricted area, suspects attempting to hide where they had been will be forced to fabricate a story. On the other hand, suspects who are describing truthfully an event that did in fact occur will retrieve a description of this event from memory.

We believe that the key to distinguishing truthful dyads from deceptive dyads is the concept of transactive memory . Two people describe an event differently if they had actually performed that event together versus if they did not but are fabricating a story about an event that did not take place.

If you and I share some experience, say that we go on a fishing trip together, we encode memories of that event between us. That is, you store in memory some parts of the event, and I store some parts. When we are questioned about the event, we recall it also in a joint manner—you recall some information, and I recall some information. Our style of interaction when recalling this jointly experienced event is collaborative—we elaborate each other's sentences and fill in stories for one another. In short, information that is encoded transactively is retrieved in an interactive manner. This is not as evident when two people recall a story that is fabricated or that did not take place.

We conducted an empirical study in which two police officers/firefighters who had served together as partners took part at a time. In the “truth” conditions, they were asked to simply describe an event or call that they had been on in the recent past. The experimenter conducted an interview with each pair of officers, asking them to describe this event, their actions, their partner's role, and actions taken to resolve the problem. In the “deception” conditions, the officers were instructed to fabricate a story on the spot that did not take place, but to make the story as realistic and believable as possible. We expected that the officers in the truth-telling dyads would be able to draw on shared memory of the actual event they had participated in and would be able to describe this event in an interactive manner. In contrast, deceptive dyads are forced to fabricate an event (an individual-level cognitive task) and are more likely to describe that event from an individual perspective. Results bore this out. In contrast to those in truth-telling dyads, those in deceptive dyads were less likely to conduct back-and-forth exchanges, less likely to look at one another, and less likely to exhibit synchrony in communication ( the degree to which one member of the dyad exhibits similar behaviors and uses similar linguistic styles as the other member). Moreover, these effects were substantial.

Study 2

Viewing deception from a social or group perspective led us to consider another side of the coin: how the “power” of teams may be exploited by using two interviewers in an interview setting. For example, we have the adage that two heads can often be better than one, but it has also been lightheartedly noted that the optimal size for a group is 0.7 persons.

In initially considering this question, we made two observations. First, tandem interviewers are used with some regularity in interviewing suspects in law enforcement settings. Somewhat surprisingly, we were not able to discern any specific rationale for the use of one or two interviewers. Officers questioned stated that it did not depend on any specific factor, such as the severity of the crime, but usually on whether another case officer was present and available for the interview. Second, concern was voiced in the literature that employing two interviewers may have negative consequences, specifically that “three is a crowd” and that it may be “more difficult to establish rapport” (Defense Academy for Credibility Assessment, 2007, pp. 2–12). Moreover, the Department of Defense field manual on intelligence collection cautions that a third person in the communications loop may negatively impact the establishment of rapport (Department of the Army, 2006). Therefore, we conducted an initial study to examine the possible negative effects of a third party on rapport in the investigative interview (T. Driskell, Blickensderfer, & Salas, 2013).

Investigative interview transcripts acquired from state and local police departments in the southeastern United States were used as the source material for this study. We requested that these agencies provide samples of investigative interview transcripts that had been adjudicated and were available in their files. Of the 55 transcripts obtained, 30 were dyadic interviews (a single interviewer and interviewee), and 25 were triadic interviews (two interviewers and an interviewee). We conducted a linguistic analysis of the interview content to examine three components of rapport: (a) mutual attentiveness or verbal immediacy, (b) positivity or positive affect, and (c) coordination or synchrony of interaction. The results indicated that there was no negative impact on rapport attributed to tandem interviewers.

We are currently examining the potentially more interesting question of the advantages offered by the use of teams —two individuals working together— in conducting the investigative interview. We believe that tandem interviewers can benefit from the traditional advantages of teams—the ability to pool resources, exchange information, coordinate actions, catch errors, and share responsibility for decisions. Furthermore, the task of an investigative interviewer is a demanding one, involving controlling the interview, formulating questions, evaluating responses, monitoring interviewee behavior, and framing follow-up probes, among other tasks, placing a considerable burden on the capacity of the individual interviewer. If two interviewers are employed, the intricacies and demands of the interview can be shared between them, potentially reducing task demand and freeing up cognitive resources. Furthermore, having two interviewers allows each to follow-up on the other's questions, fill in gaps, and check anomalies or inconsistencies that may not be caught by a single interviewer. In brief, the extensive research on teams would lead us to believe that teams can provide distinct advantages in the investigative interview, although these questions are untested.

Summary

Adopting a social dynamics perspective on the intelligence interview results in several consequences. First, there are unique research questions that arise, as noted above, that are not addressed from an individual-level approach. Although the term “investigative interview” may evoke the mental image of a hard-nosed interrogator on one side of the table and a sweating suspect on the other, in today's national security environment, information-gathering interviews are also likely to take place in field settings, such as checkpoints, street corners, and daily patrols. However, the predominant research perspective on the investigative interview has been at the individual level, which has resulted in less emphasis on the group dynamics of the interview context and how team-level strategies can enhance information collection in an intelligence interview environment. We propose that the gains that have been achieved from “harnessing the power of the team” in group decision making and other contexts can lead to new strategies and approaches to enhance the intelligence interview.

Acknowledgments

This research was funded by the Department of Defense. We acknowledge the contributions of Susan E. Brandon (Federal Bureau of Investigation) and Ruth Willis (Naval Research Laboratory).

References

Defense Academy for Credibility Assessment. (2007). PCASS operators course manual. Ft. Jackson, SC: Author.

Department of the Army. (2006). FM 2-22.3: Human intelligence collector operations . Retrieved from https://www.fas.org/irp/doddir/army/fm2-22-3.pdf

DePaulo, B. M., Lindsay, J. J., Malone, B. E., Muhlenbruck, L., Charlton, K., & Cooper, H. (2003). Cues to deception. Psychological Bulletin, 129, 74–118.

Driskell, J. E., Salas, E., & Driskell, T. (2012). Social indicators of deception. Human Factors , 54 , 577–588.

Driskell, T., Blickensderfer, E. L., & Salas, E. (2013). Is three a crowd? Examining rapport in investigative interviews. Group Dynamics: Theory, Research, and Practice , 17, 1–13.

Hazlett, G. (2006). Research on detection of deception: What we know vs. what we think we know. In R. A. Fein (Ed.), Educing information. Interrogation: Science and art . Intelligence Science Board Phase I Report (pp. 45–61). Washington, DC: National Defense Intelligence College.

Jackson, B. A. (2007, January–February). Counterinsurgency intelligence in a “long war”: The British experience in Northern Ireland. Military Review, 74–85. Retrieved from http://www.rand.org/content/dam/rand/pubs/reprints/2007/RAND_RP1247.pdf

Miller, A. (2005, January 20). “Roughriders” nab Latifiyah bombmaking cell. Defend America: U.S. Department of Defense News About the War on Terrorism. Retrieved from http://www.defendamerica.mil/articles/jan2005/a012005wm2.html

Mullen, B., Bryant, B., & Driskell, J. E. (1997). The presence of others and arousal: An integration. Group Dynamics, 1, 52–64.

Table 1

Existing Research on Deception

		Focus of observation or interview
		Individual	Joint
Level of analysis	Individual	A Deception research	B ----------
	Joint	C Liars in collusion	D Conspiracy