Task Performance Under Deceptive Conditions: Using Military ...

57 downloads 908 Views 264KB Size Report
support for participants in games using a face-to- face communication method to have lower perceptions of deception and task difficulty when compared to ...
Proceedings of the 38th Hawaii International Conference on System Sciences - 2005

Task Performance Under Deceptive Conditions: Using Military Scenarios in Deception Detection Research

David P. Biros [email protected] Michael C. Hass Air Force Institute of Technology

Abstract The goal of this research was to investigate how changes in modality (communication type) and external conditioning (warnings of player deception) relate to perceptions of deception and task difficulty and, in turn, how these perceptions relate to the final group game scores in a cooperative effort with conflicting goals. One hundred and eight participants were grouped into teams of three, given similar instructions but different goals, and asked to play a cooperative game called StrikeCOM that simulates the intelligence gathering needed to develop an air tasking order and subsequent air strike on three military targets. The analysis of the post-game surveys showed support for participants in games using a face-toface communication method to have lower perceptions of deception and task difficulty when compared to games using real-time plain text chat.

1.0 Introduction Deception is part of everyday life [8, 21]. Examples of this range from the frivolous, such as agreeing that a style of hair is beautiful when you feel that it is not, to the serious, such as courtroom testimony, to the life-critical, which can occur during military conflict. Despite this inundation, it has been found that people are typically poor detectors of deception- commonly only able to detect it at a level slightly better than chance [9, 18]. Why people are typically so poor at detecting deception communication is apparent when one considers the nature of communication and of people.

Karl Wiers, Douglas Twitchell, Mark Adkins, Judee K. Burgoon, Jay F. Nunamaker Jr. University of Arizona

The basic nature of communication is to convey information from sender to receiver through some active means. This means that when there is communication, the receiver is attempting to comprehend what the sender is saying and there is a basic assumption made that the message is comprehensive and truthful [12]. The problem with this is that research has shown that such a mindset can lead to truth bias; a predisposition to assume that all others’ communication is truthful or trustworthy [13, 17]. In a military environment, it is imperative for members to be able to trust each other when making critical decisions in support of mission accomplishment. For complex tasks, multiple team members are often ;brought together to analyze data, gain situational awareness, and develop optimal courses of action to complete a mission. An example of this is the Air Operations Center (AOC). It is the mission of the AOC to determine the Air Tasking Order (ATO). Multiple team members from various military disciplines determine the operation mission (flying mission) for the war-fighters. Information integrity is critical. The introduction of deception into such an environment could be detrimental thus, it is imperative that AOC crew members remain vigilance to the possibility of deception. This study develops deception and deception detection models by examining group performance and perceptions of deception and task difficulty under two different media types or modalities commonly employed in military campaigns and two different levels of awareness using a military-based scenario. The two media types considered are face-to-face communication and real-time text chat. The two levels of awareness are manipulated through the introduction of additional information to selected participants which may make them more

0-7695-2268-8/05/$20.00 (C) 2005 IEEE

1

Proceedings of the 38th Hawaii International Conference on System Sciences - 2005

suspicious of the other group members. The scenario developed was one created using a software package called StrikeCOM, developed by the Center for the Management of Information at the University of Arizona to evaluate group performance in a task requiring a coordinated effort among players.

to e-mail, video conferencing, and chat rooms [2, 10]. Given this increasing emphasis on technologically-based communication, the probability for deceit within this media increases [25]. Thus, it is appropriate to consider the influence of deception under different modalities.

2.0 Background

2.1 Modality

There has been a significant amount of attention paid to the field of deception research, and several theories and models have been presented. One of the more significant of these is Interpersonal Deception Theory (IDT) [4] The theory was developed to identify the characteristics of deceptive communication between a deceiver and one or more receivers [4]. It takes into account the dynamic nature of communication, where participants may modify their style of communication based on the feedback they receive. The IDT relies on a two-part definition of interpersonal communication and deceptive communication to establish the theory scope. Interpersonal communication is defined as the “dynamic exchange of messages between two (or more) people” [4: 205]. This dynamic exchange requires that the sender and receiver are active participants in the communication and that individual roles will change over time, as communicators become listeners and vice versa. Another reason why people have difficulty detecting deception has to do with their preconceptions of what are accurate cues to deception. So which cues do people associate with deception? Surveys have shown that most people link gaze aversion and fidgeting with deception [1, 14, 24]. In one survey, 75 percent of police officers believed that liars look away. One possible reason for this is that the police manuals on interrogation promote this idea even though there is little evidence to back this up [11]. These inaccurate preconceptions make detecting deception more difficult. Two recent studies that examine the relation between what people think are associated with deception and their ability to detect it have shown this apparent conflict. Police officers that believe that liars avert their gaze and fidget were shown to be among the worst at detecting deception [15, 23]. Only when the police officers were asked to review the video tapes for specific cues did the detection success rates increase. Another facet to this issue is that changes in technology has made face-to-face and telephone conversations to be used less often when compared

Modality refers to the different communication media or modes that can be employed (face-to-face, e-mail, telephone, etc) when sending information to one or more recipients. These media have different characteristics that affect how they convey information, how much information each can convey, and how many different people can they convey information to in a set amount of time [4, 6, 7, 20]. Varying modalities allows the examination of different sets of deceptive indicators and as well as the influence of media richness on the ability to detect deception. Along with varying modalities, decision making behaviors of individuals and, in turn, of groups may be influenced by the presences of external conditioning. This may be in the form of increased awareness of a potential deceptive act taking place or it may be in some other form.

2.2 External Conditioning External conditioning is the presence of information provided from an outside source to certain individuals or group members about the possibility of deception The goal of providing this information is to raise the non-specific suspicion levels of certain group members by providing an external stimulus to observe individual changes in perception of deception and task difficulty [19]. A previous study has determined that external stimulation or warnings are positively associated with deception detection success [3] and the purpose of including this condition is to expand these results to consider its interaction with modality.

3.0 Research Model In order to capture the influences on medial type and external conditioning on deception detection ability we offer the following conceptual model (Figure 1). Varying media types or modalities should influence ones perceptions of deception and task difficulty. Similarly, the

0-7695-2268-8/05/$20.00 (C) 2005 IEEE

2

Proceedings of the 38th Hawaii International Conference on System Sciences - 2005

presence of external conditioning may also influence in human perception in a similar manner.

These perceptions, in turn, will affect task success (i.e. deception detection success).

Figure 1: Media Type and External Conditioning Model

The concept of media richness [6] suggests that as communications media increases in capability, the quality of the interaction between two the sender and receiver improves. When performing decision-making tasks, varying the modality of the communication in a deceptive environment should influence the perception of deception of the decision-maker. Further, this change in modality should also influences and individuals perception of difficulty in a decisionmaking task. That is, the less rich communication modes are likely to make a decision-making task seem more difficult that those that are richer. In turn, higher levels of media richness should then result in greater level of task performance. As such we posit the following hypotheses: H1a: Tasks performed using a text-only communication method will have a higher perception of deception when compared to tasks performed using a face-to-face communication method.

H1b: Task performed using a face-to-face communication method will be perceived as easier to perform when compared to tasks performed using a text-only communication method. H1c: Task scores will be higher on average for those employing the face-to-face communication method when compared to those using the faceto-face communication method. However, modality alone may not be enough. Understanding the role of situational awareness is also necessary. Individual who have a higher awareness of the presence of deceptive in formation may react differently under varying modalities. In fact, O’Hair and Cody [19] suggest that external stimuli such as a warning of the presence of deception may result in greater levels of success at deception detection tasks. However, that same warning may result in higher levels of perceived task difficulty.

0-7695-2268-8/05/$20.00 (C) 2005 IEEE

3

Proceedings of the 38th Hawaii International Conference on System Sciences - 2005

H2a: The presence of external conditioning is associated with a higher perception of deception. H2b: The presence of external conditioning is associated with a higher perception of task difficulty. The perception of deception should have an influence on deception detection task performance. Those who believe there is deception present should be more vigilant and thus look harder for the deceptive information resulting in better task performance. This was demonstrated by in an earlier study [3], but the influence of its interaction with modality is yet unknown. Additionally, those who feel the task is difficult are likely to performance less well [3]. As such, once modality and external conditioning influence perceptions of task difficulty and the presence of deception, the will ultimate influence task success [26] H3: A higher perception of deception is associated with higher average task performance. H4: A higher perception of task difficulty is associated with lower average task performance. In summary, varying media types and external conditioning should influence an individual’s perception of deception and perception of task difficulty. When these two perceptions are affected, this, in turn, can influence individual and group task success. In the next section we introduce a novel method for testing our hypotheses.

4.0 Experiment In order to test these hypotheses and study the influence of media type and external conditioning, we developed an experiment whereby three individuals were teamed to solve of problem. A common problem for Air Force officers to solve is to determine which targets aircraft should be direction to in order to achieve military objectives on the battlefield. In the real

world, this activity in performed in an Air Operations Center (AOC) and a group of military personnel including coalition forces, team to determine the target priorities for the their aerospace resources. Members of the AOC provide information and subject matter expertise to help the group decide on an optimal list of target priorities. This prioritized list is referred to as an Air Tasking Order (ATO). Thus, in order to test our hypotheses, we developed a simulated AOC environment and required our decision-makers to devise an ATO. We did this using an AOC simulator called StrikeCOM.

4.1 StrikeCOM StrikeCOM is a game where teams of two or more players cooperate to determine three targets that are hidden over a 6x6 grid map (See Figure 4). Each player is given assets that they could use once per turn. For this study, we used three player teams with each player providing information about two information assets (e.g. air, space, human intelligence, etc). The two assets had different search coverage abilities; asset one could search three grid squares per turn and asset two could search one grid square per turn. Search efforts encompassed five rounds where each person used their assets to search different portions of the map for possible targets. Results of each search yielded information about the grids searched. Each grid searched showed that it either had no target, possibly had a target, or probably had a target. Conducting another search on a grid that possibly had a target would have shown if there was either no target or probably a target there. Due to the number of grids on the map, it was impossible for any one of the players to search the entire map by themselves. Only the individual players knew the results of their search. They needed to communicate their search results to the other players in order to develop a game-winning or optimal strategy. In order to have the greatest chance of finding targets, players had to plan and coordinate their searches using the communication mode they were provided. On the sixth and final round, each player selected a set of three or more grids to attack in the hopes of destroying the three targets (see Figure 2).

0-7695-2268-8/05/$20.00 (C) 2005 IEEE

4

Proceedings of the 38th Hawaii International Conference on System Sciences - 2005

Figure 2. View of StrikeCOM

The number of group strike selections that correctly chose the correct target locations determined the final game score. A perfect score was achieved when all group members selected the same three correct targets for attack. This game is made more difficult in this experiment by the fact that one of the three players does not want targets to be found or destroyed and will likely provide misleading information to the other two players. In short, one player was told insert deceptive information into the decision making task.

One of the advantages of StrikeCOM is that it is an easy game to learn. It was even easier for the cadets as they had a familiarity with the AOC environment. After a short introduction and practice period, the cadets were ready to begin the task of determining the ATO. Each group of cadets was expected to complete their session within 2 hours and the room for the experiment allowed for up to two simultaneous groups. Each participant was videotaped for the duration of the session. All audio and text inputs were recorded and transcribed for future analysis.

4.2 Experimental Design To conduct the experiment, we recruited cadets from a Air Force Reserve Officer Training Corps (ROTC) detachment located at a university in the southwestern United States. In their ROTC curriculum, cadets learn about AOCs and ATOs. The detachment commander believed that our experiment would be a great opportunity for the cadets. In all, we recruited 108 cadets. The cadets were randomly assigned in groups of three (36 groups) and given positions and information assets for the AOC simulator, StikeCOM.

4.3 Independent Variables: Role, Deception, and External Suspicion Induction Each player was selected to play the role of one of three component commanders: Air, Intel, and Space. Each component had a different role within the game and participants were randomly selected for each role at the beginning of the game. The Air component commander was given the basic set of instructions. They were told how to play the game and their goal is to play the game as

0-7695-2268-8/05/$20.00 (C) 2005 IEEE

5

Proceedings of the 38th Hawaii International Conference on System Sciences - 2005

best they can and help to achieve the highest overall group score. They were not made aware that any other player might have had a different goal. For classification purposes, the Air component commander was considered the naïve player. The Intel component commanders were also given the same basic set of instructions but half of them were also given one additional set of instructions They were provided with an external source of suspicion by being informed that one of the other two players may provide deceptive information. The Intel component commander did not know which of the other players was the deceiver and may have tried to find out whom though their goal remained to help achieve the highest overall group score in the game. The Space component commanders were given a similar set of instructions, however, unlike the other two, the space commanders were told to correct location of the targets. They were also told to lead the other two players away from the correct targets using any means necessary.

measurements were found to be reliable in a previous study. [28] All questions used in directly answering the hypotheses used the same scale ratings of 1 (strongly disagree) to 7 (strongly agree).. A factor analysis was performed to ensure the questions loaded on the deception and task difficulty constructs. The results of the factor analysis show that some of the variables in both the deception and task difficulty measures are similar. A rotated component matrix depicted that the questions were appropriately similar to be combined into a composite score to evaluate the perception of deception. An additional factor analysis was accomplished for the perception of task difficulty questions. Four out of the five questions in the group loaded together under the task difficulty construct. The one question that did not load had to do with functions of the game instead of issues concerning the task. The measurement for task success was calculated by the game as described earlier.

6.0 Analysis 5.0 Measurement In order to successfully test the hypotheses posited for perception of deception and perception of task difficulty, the groups were divided into two modality types: face-to-face and text-chat. In accordance with Daft and Lengal, [6] we consider face-to-face a richer modality that text-chat. We also divided the groups by level of external conditioning (Air had no external conditioning, Intel received external conditioning). Because the participants that played the role of Space component commander had direct knowledge of the locations of the enemy camps and were instructed to deceive the other members, their perceptions of deception and task difficulty would be different from the other team members and are excluded from the analysis of post-game survey data. Measurements of the perception of deception were obtained by having the participant answer a questionnaire that was directly related to evaluating the level of suspicion the individual participant had of their team members and their belief that their team members may have been deceitful. This questionnaire was administered at the end of play so as not to influence the participants. Measurements of the perception of task difficulty were obtained though analysis of the questions from a “task difficulty’ measure. Both

Once the data collection was complete, we performed a pair of factorial ANOVAs to test for hypothesis support while taking into account the possibility of an interaction effect between modality (face-to-face) and external conditioning (Intel and Air) while examining the perceptions of deception and task difficulty. The results of the factorial ANOVAs (Į = 0.05) using a one-tailed analysis show that there is no significant interaction between modality and external conditioning for either perception of deception (Fratio = 1.664, observed significance = 0.203) or perception of task difficulty (F-ratio = 1.541, observed significance = 0.22). This enabled us to consider modality and external conditioning as not having a joint , yet be able to analyze their interactive affects.

6.1 Analysis of Hypotheses Hypothesis H1a stated that tasked performed using a text-only communication method will have a higher perception of deception when compared to games performed using a faceto-face communication method. An ANOVA (all ANOVAs performed at Į = 0.05 using a one-tailed analysis) indicated that the perception of deception scores were higher for text-only games when compared to face-to-face games (mean = 3.85 TXT

0-7695-2268-8/05/$20.00 (C) 2005 IEEE

6

Proceedings of the 38th Hawaii International Conference on System Sciences - 2005

and 3.05 FTF) and that the difference is significant (F-ratio = 4.44, observed significance = 0.04). Participant playing StrikeCOM under the text-only communication method perceived the presence of deception to a greater extent than the games where participants communicate face-to-face. Hypothesis H1b posited that the tasked performed using a face-to-face communication method will be perceived as easier to perform when compared to games performed using a textonly communication method. The results show that the perception of task difficulty was higher for text only games when compared to face-to-face games (mean = 3.96 TXT and 2.87 FTF) and that the difference is also significant (F-ratio = 8.97, observed significance = 0.004). Participants using the StrikeCOM games felt that the face-to-face tasks were much easier to accomplish when compared to the text-only games. Hypothesis H1c stated that the final group game scores will be higher on average for those employing the face-to-face communication method when compared to those using the text-only communication method. One would expect a higher level of media richness to enable them to perform the task better. However, even with the external conditioning considered, the average game scores for text-only and face-to-face games are almost identical (mean = 0.203 TXT and 0.200 FTF) and there is no significant difference between them (F-ratio = 0.005, observed significance = 0.944). The face-to-face and text-chat task performance measures were quite close in comparison. Hypothesis H2a maintains that the presence of external conditioning is associated with a higher perception of deception. The factorial ANOVA results (performed at Į = 0.05 using a one-tailed analysis) show that participants who received an external warning of the possibility of player deception (Intel participants) had higher perception of deception than those who did not receive any warning (Air participants) (mean = 3.81 Intel and 3.08 Air) regardless of what type of StrikeCOM game was played. While the difference is not significant (F-ratio = 3.68, observed significance = 0.06), the results are strong enough to suggest continued study of the hypothesis. This result is not overly surprising. Past studies have demonstrated that users will continue to rely on information technology output even when its veracity was in question [26, 27] Hypothesis H2b posited that the presence of external conditioning is associated with a higher perception of task difficulty. One would expect the introduction of the idea that some information is

deceptive would increase the cognitive task load of the player. The results indicated that the perception of task difficulty is slightly higher on average in Intel participants when compared to Air participants (mean = 3.47 Intel and 3.35 Air) but this difference is not significant (F-ratio = 0.100, observed significance = 0.753). The Intel participants (those with external conditioning) may have found the task more difficult but that the difference is too small to say that for certain.

6.2 Analysis of Effects of Perceptions on Task Scores The analysis of the effect of perceptions of deception and task difficulty on the final group game scores was performed using linear regression (Į = 0.05). Hypothesis H3 stated that a higher perception of deception is associated with higher average game scores. Those groups with Intel commanders who received the external conditioning would likely counter the planted deception. The regression results show a strong negative relationship (bivariate fit: Game Score = 0.276248 - 0.0215227 Perception of Deception) between perception of deception and group game score (F-ratio = 8.26, observed significance = 0.0046). This means that the alternate of H3, that a higher perception of deception is associated with lower game scores, is supported rather than the original hypothesis and means that, in general, as the individual perception of deception increased, the final StrikeCOM group game score decreases. This result will be discussed later on. Hypothesis H4 maintain that a higher perception of task difficulty is associated with lower average game scores. The regression results show a weak negative relationship (bivariate fit: Game Score = 0.2076717 - 0.0017208 Task Difficulty) between the perception of task difficulty and the final group game score. This weak relationship is not significant (F-ratio = 0.078, observed significance = 0.78) and H4 cannot be supported. Thus, increasing individual perception of task difficulty had no significant effect on the final StrikeCOM group game score. The level of perceived difficult apparently had no affect on task performance in this case.

7.0 Discussion Collectively, hypotheses H1a, H1b, and H1c proposed that changes in modality would have a significant effect on the on the perceptions of

0-7695-2268-8/05/$20.00 (C) 2005 IEEE

7

Proceedings of the 38th Hawaii International Conference on System Sciences - 2005

deception and task difficulty and on the final group game scores. The statistical tests support the changes in modality affecting individual perceptions but not affecting the task performance. An attempt to explain why there was no difference in mean game score between modalities requires a reexamination of the key differences in media characteristics as illustrated by Carlson [5] in Chapter 2 between the face-to-face and text-only games. The two media types would be similar in terms of symbol variety and tailorability. The media would also be similar in terms of reprocessability due to the presence of scratch paper (which all players used) in all games providing the ability to make written logs of results and suggestions. Face-to-face games would provide a slightly higher synchronicity (by a few seconds) and conversely text-only games would provide a slightly higher level of rehearsability. The biggest difference between the two media types is in the area of cue multiplicity where faceto-face games would be able to provide visual, verbal, and nonverbal cue channels while text-only games provide a verbal (plain text) cue channel only. Additionally, we observed that the textonly games took significantly longer to complete compared to face-to-face games (on the order of twice as long). This is understandable because it can be expected to take longer to communicate a complex idea using typed plain text compared to a face-to-face conversation. It can be noted however that while the text-only games took longer to complete, the research team allowed the participants uninterrupted time to complete the games even when their games ran over into the next study time slot. This could mean that, given enough time to communicate ideas within a group, the difference in channel cues, in verbal and nonverbal communication, may not have enough of an effect to change the final outcome. If, however, we held the text only group to the time allotted, their overall scores may have been different. We believe this to be a limitation in the study.

7.1 External Conditioning Collectively, hypotheses H2a and H2b proposed that the presence or absence of external conditioning would have an effect on the individual perceptions of deception and task difficulty. Analyses of these hypotheses provided limited support at best but did show the potential for support if this presence of external conditioning is coupled with a media type with low cue

multiplicity such as text-chat or voice. The results of studying external conditioning versus perception of deception provide a limited reinforcement to a previous study that found support to the idea “that warnings about possible deception in computerbased data will be positively associated with detection success” [3: 14]. Future studies could examine the interactions between modality, external conditioning, and training in order to expand on the work performed here and in other studies [2].

7.3 Individual Perceptions and Task Performance Hypotheses H3 and H4 were developed to examine what effect individual perceptions of deception and task difficulty had on game score. The analyses of these hypotheses show that a greater individual perception of deception can be associated with a lower average group game score, however, there is no correlation between perceptions of individual task difficulty and group game score. Results from this study reinforce the idea that media characteristics and external conditioning can affect deception detection accuracy. These results are beneficial to the understanding of interactive deception and deception detection processes from the view of the academic and the practitioner. The lessons learned and consequences stemming from the discoveries and limitations identified in this and the preceding chapter can be applied to future studies in the hope of further increasing the pool of knowledge on interactive deception processes.

8.0 Conclusions The influence of modality and external conditioning on task performance under deceptive conditions is indeed and interesting phenomenon. Varying the modality changes individual perception about the difficulty of the task, but seems to have little effect on task performance. The affect of external conditioning or warning seemed to be minimal at best. External warnings to some group members regarding the potential for deception did little improve task performance when deception was present. This underscores the significant of the truth bias construct. The use of the StrikeCOM game provided a unique and interesting why to examine the influence of modality and external conditioning on

0-7695-2268-8/05/$20.00 (C) 2005 IEEE

8

Proceedings of the 38th Hawaii International Conference on System Sciences - 2005

task performance. It provided the respondents with a simulated environment that not only helped them to understanding the function of an AOC, but also helped to increase respondent motivation to succeed at the task. Continue studies using this tool under varying conditions are indeed warranted.

References 1.

2.

3.

Akehurst, L., G. Kohnken, A. Vrij, and R. Bull, “Lay Persons’ and Police Officers’ Beliefs Regarding Deceptive Behavior,” Applied Cognitive Psychology, 10: 461-471 (1996). Biros, David P. The Effects of Truth Bias on Artifact-User Relationships: An Investigation of Factors for Improving Deception Detection in Artifact Produced Information. PhD Dissertation. Florida State University, Tallahassee, FL, 1998. Biros, David. P., Joey F. George, and Robert W. Zmud. “Introducing Sensitivity to Deception in Order to Improve Decision Making Performance: A Field Study,” MIS Quarterly 26:1-26 (June 2002).

4.

Buller, David B. and Burgoon, Judee K. “Interpersonal Deception Theory,” Communication Theory, 6: 203 – 242 (August 1996).

5.

Carlson, John R., Joey F. George, Judee K. Burgoon, Mark Adkins, and Cindy H. White, “Deception in Computer-Mediated Communication,” To appear in a special issue on deception of Group Decision and Negotiation (2004).

6.

7.

8.

9.

Daft, Richard L. and Robert H. Lengel, “Organizational Information Requirements: Media Richness and Structural Design,” Management Science, 32: 554-571 (May, 1986). Dennis, Alan R and Joseph S. Valacich, “Rethinking Media Richness: Towards a Theory of Media Synchronicity,” Proceedings of the 32nd Hawaii International Conference on Systems Sciences, Maui, HI (1999). DePaulo, Bella M., D. A. Kashy, S. E. Kirkendol, M. M. Wyer, and J. A. Epstein, “Lying in Everyday Life,” Journal of Personality and Social Psychology, 70: 979-995, (1996). Feeley, Thomas H. and Mark A. deTurck, “Global Cue Usage in Behavioral Lie Detection,” Communication Quarterly, 46: 109-131 (Fall 1995).

10. George, Joey F. and John R. Carlson, “Electronic Lies: Lying to Others and Detecting Lies Using Electronic Media,” Proceedings of the Fifth Americas Conference on Information Systems. 612614. Milwaukee, WI, (August 1999a). 11. Gordon, N. J., and W. L. Fleisher, Effective Interviewing and Interrogation Techniques. San Diego, CA: Academic Press, (2002). 12. Grice, H. P. Studies in the Way of Words. Canbridge, MA: Harvard University Press, 1989. 13. Levine, Timothy R. and Steven A. McCornack, “Linking Love to Lies: A Formal Test of the McCornack and Parks model of Deception Detection,” Journal of Social and Personal Relationships, 9: 143 – 154 (1992). 14. Mann, S. “Suspects, Lies, and Videotape: An Investigation into Telling and Detecting Lies in Police / Suspect Interviews,” Unpublished PhD thesis. University of Portsmouth (United Kingdom), Psychology Department, 2001. 15. Mann, S., A. Vrij, and R. Bull, “Detecting True Lie: Police Officers’ Ability to Detect Suspects’ Lies.” Manuscript submitted for publication. 16. McCornack, Steven A. “Information Manipulation Theory,” Communication Monographs, 59: 1-16 (March 1992). 17. McCornack, S. A., and M. Parks, “Deception Detection and the Other Side of Trust,” Communication Yearbook, 9: 377-389, (1986). 18. Miller, G. R., and J. B. Stiff, Deceptive Communication. Sage Publications, Inc, 1993. 19. O’Hair, Dan H. and Michael J. Cody. The Dark Side of Interpersonal Communication. New Jersey: Lawrence Erlbaum Associates, 1994. 20. Short, J., E. Williams, and B. Christie, The Social Psychology of Telecommunication, London: John Wiley and Sons, 1976. 21. Turner, R. E., C. Edgley, and G. Olmstead, “Information Control in Conversations: Honesty is Not Always the Best Policy,” Kansas Journal of Speech, 11: 69-89 (1975). 22. Vrij, A. Detecting Lies and Deceit: The Psychology of Lying and the Implications for Professional Practice. Chichester: Wiley and Sons, 2000. 23. Vrij, A., and S. Mann, “Telling and Detecting Lies in a High-Stake Situation: The Case of a Convicted Murderer,” Applied Cognitive Psychology, 15: 187203 (2001).

0-7695-2268-8/05/$20.00 (C) 2005 IEEE

9

Proceedings of the 38th Hawaii International Conference on System Sciences - 2005

24. Vrij, A., and G. R. Semin, “Lie Experts’ Beliefs about Nonverbal Indicators of Deception,” Journal of Nonverbal Behavior, 20: 65-80 (1996). 25. Zmud, R. W. “Opportunities for Strategic Information Manipulation Through New Information Technology,” in Organizations and Communications Technology, J, Fulk and C. Steinfield Eds. Newbury Park, CA: Sage, 95-116 (1990).

27. Muir, B.M. and Moray. N., (1996) “Trust in automation: Part II. Experimental studies of trust and human intervention in a process control simulation,” Ergonomics, 37(11): 429-460. 28. Burgoon, Judee K., J. Bonito, and K. Kam, “Communication and Trust Under Face-to-face and Mediated Conditions: Implications for Leading from a Distance.” In Leadership at a Distance. Ed. S. Weisband and L. Atwater. Mahwah, NJ: LEA, in-press.

26. Biros, D.P., Daly, M. A., and Gunsch G., “Task Load and Automation Use in an Uncertain Environment,” Group Decision and Negotiation, 13 (2), March 2004.

0-7695-2268-8/05/$20.00 (C) 2005 IEEE

10