Anger as a cue to truthfulness

ARTICLE IN PRESS Journal of Experimental Social Psychology xxx (2010) xxx–xxx

Contents lists available at ScienceDirect

Journal of Experimental Social Psychology journal homepage: www.elsevier.com/locate/jesp

FlashReport

Anger as a cue to truthfulness Jessica L. Hatz *, Martin J. Bourgeois Florida Gulf Coast University, Fort Myers, FL, United States

a r t i c l e

i n f o

Article history: Received 28 December 2009 Revised 22 February 2010 Available online xxxx Keywords: Deception Emotion Anger

a b s t r a c t A fairly robust finding in the deception literature is that lie-tellers show more negative emotion than truth-tellers. Ekman (1985), however, has reasoned that a specific type of negative emotion – anger – is especially difficult to feign and therefore should be more prevalent in truth-tellers who are falsely accused of a transgression than in lie-tellers who are guilty. To our knowledge, Ekman’s prediction has not yet been empirically tested. By comparing the verbal and nonverbal cues associated with truths and lies across a number of lie-eliciting situations, we demonstrate that truth-tellers accused of a wrongdoing do show more anger, both verbally and nonverbally, than lie-tellers accused of the same act, but only in situations where students choose to commit a transgression (or not) and actually believe themselves to be in trouble. Results underlie the importance of taking into consideration the type of lie being told in order to accurately predict deceptive cues. Ó 2010 Elsevier Inc. All rights reserved.

Introduction An overwhelmingly robust finding in deception research is that negativity is a reliable cue to deception (DePaulo et al., 2003; Vrij, 2000). In perhaps the most comprehensive examination of this issue, DePaulo et al. (2003) conducted a meta-analysis examining the cues that distinguish liars from truth-tellers. Consistent with DePaulo et al’s. self-presentational perspective, liars across numerous studies were more negative and less cooperative than truthtellers. Specifically, liars, compared to people telling the truth, consistently showed more fear, nervousness, and hostility; and this was especially true when people were highly motivated to get away with their lies and when people told lies about having committed a transgression. It would appear that a reliable and efficient way to distinguish truth-tellers from liars would be to assume that any negativity present in a speaker’s words or demeanor is indicative of deceitfulness. Indeed, people often use negative behaviors, such as nervousness, fidgeting, and fear, as cues to deception (Vrij, 2000). Furthermore, the use of the polygraph to detect deception relies on similar logic; that is, if someone is lying, it is believed that this will be revealed through the increased sympathetic nervous system arousal that accompanies negative affect (Iacono, 2008). However, Ekman (1985) has reasoned that a specific type of negative emotion – anger – should actually be more prevalent in truth-tellers who are falsely accused of a crime than in lie-tellers who are guilty (Ekman, 1985). Presumably, a person accused of a transgression should feel angry and indignant at this false accusation. Because the muscle * Corresponding author. E-mail address: [email protected] (J.L. Hatz).

movements responsible for negative expressions such as fear, disgust, and anger are difficult to control, an individual who is wrongly accused will find it difficult to conceal their anger, just as a guilty suspect will find it difficult to feign the angry expressions they associate with wrongful accusations (Ekman, 1985). To our knowledge, Ekman’s prediction has not yet been empirically tested; indeed, anger was not even studied as a possible cue in the Depaulo et al. (2003) meta-analysis of cues to deception. Empirical examination would require the elicitation of the types of lies told in real-world, high-stakes settings – the liar must believe he/she is actually being accused of a transgression, and must also be motivated to convince the accuser of their innocence. This type of lie is difficult to elicit in laboratory settings. In fact, Vrij (2004) has argued that ‘‘laboratory studies are not suitable for examining the responses in high-stake situations, as raising the stakes really high is not possible in laboratory settings for ethical reasons” (p. 289). Therefore, the studies in which negativity signaled deceit rather than truthfulness typically have utilized lower-stake situations in which truthful participants likely experienced little (if any) discomfort, and therefore would not have had any reason to become angry at a wrongful accusation. Several of these experiments, for example, utilized a mock crime paradigm, in which participants are instructed to commit a transgression, such as hiding a piece of evidence from an interrogator or stealing money, and then further instructed to claim that they are innocent (see Frank & Ekman, 1997, 2004; Vrij, 1995). Thus, it may be that the lies and truths generated in these studies are somewhat different than those experienced when people are accused of actual transgressions, in that these participants neither chose to commit a transgression nor chose to lie about it.

0022-1031/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.jesp.2010.02.009

Please cite this article in press as: Hatz, J. L., & Bourgeois, M. J. Anger as a cue to truthfulness. Journal of Experimental Social Psychology (2010), doi:10.1016/ j.jesp.2010.02.009

ARTICLE IN PRESS 2

J.L. Hatz, M.J. Bourgeois / Journal of Experimental Social Psychology xxx (2010) xxx–xxx

The aim of the present research was to investigate Ekman’s prediction that, in contrast to the typical finding that negative emotion is associated with lying, truth-tellers should show greater anger than liars. To test this prediction, we created a situation in which participants chose whether or not to commit a transgression, and then chose whether to lie about it (adapted from Russano, Meissner, Narchet, & Kassin, 2005). We compared this situation to a condition in which participants committed a mock wrongdoing – or not – and then followed an experimenter’s instruction to either feign or truthfully maintain their innocence. Then we examined the nonverbal and verbal indicators of anger. We expected that truth-tellers would appear angrier in their nonverbal and verbal cues than lie-tellers. Nonverbal cues Methods Participants In Phase 1, 60 participants were recruited from introductory psychology. In Phase 2, 79 participants viewed videotapes of the participants in Phase 1 and rated them for expressed anger. All students earned course credit for their participation. Procedure In Phase 1 we utilized two methods for eliciting truthful and deceptive statements: real transgression (RT) and mock transgression (MT) scenarios. The RT condition utilized a novel method developed to induce normal college students to tell serious lies. In brief, participants in this method are induced to cheat (or not) on some logic problems, are accused of cheating and threatened with a substantial punishment if it was decided that they had cheated, and are then questioned about the incident while being secretly videotaped. Hence, the students chose of their own volition whether to cheat, and also chose whether to lie about having cheated. To begin the experiment, RT participants were paired with a confederate posing as another participant. After providing informed consent, the participant and confederate received oral instructions explaining the purpose of the study (ostensibly, to develop problems that would be used to place freshmen in math classes). They were told that they would be working on problems together and then individually, and they were further informed that they would be videotaped as they discussed the solutions to the team problems, but that they should not discuss the solutions to the individual problems. In the absence of the experimenter, the confederate and participant completed the team problems first, and then the individual problems. During completion of the individual problems, the confederate asked half the participants (randomly assigned to the lie condition) for help with two of the problems. Participants were thus induced to cheat. The remaining participants (randomly assigned to the truth condition) were not asked for help. Next, the confederate and participant completed a filler task while the experimenter ostensibly scored the exams. Then the experimenter, who was blind to the truth/lie condition, returned and told them that they had the same wrong answers on some of the individual problems, thus implying that they cheated. The experimenter then left the room to ‘‘contact the professor in charge of the study” about the incident. After 5 min, the experimenter returned and informed them that the professor ‘‘seemed pretty upset about what happened” and might consider this to be a case of cheating, in which case ‘‘it will be treated like any other case of academic dishonesty – you’ll go before a university disciplinary board and each be assigned an attorney.” Finally, these RT participants were secretly videotaped as they responded to the following

questions: ‘‘Did you two cheat on these problems? How would you explain that you both got the same wrong answers on the most difficult problems? Did you talk while you were working on these? Did you look at each other’s papers?,” and finally, ‘‘Do you have anything else to add?” Next, each participant was probed for suspicion and extensively debriefed. Notably, every participant given the opportunity to cheat (RT lie condition) did, in fact, cheat. Furthermore, everyone who cheated chose to lie and claimed not to have done so. Mock transgression participants also signed up for what they believed was a math study and were paired with a confederate. In this condition, participants were immediately made aware that the confederate was part of the research team and that they would be working with this person on a task. The researcher further instructed each participant, ‘‘I’m going to leave the room, and my assistant will give you some instructions about the task we’re asking you to perform. Please follow her instructions.” Next, the experimenter left (so as to remain blind to the participant’s randomly assigned lie/truth condition), and the confederate delivered the remaining instructions. Participants were told they would be taking part in a lie-detection task, and that they would be either lying or telling the truth in response to a series of questions the experimenter would ask later during the session. Next, participants randomly assigned to the lie condition were instructed to ‘‘cheat” and work through the problems with the research assistant, whereas those in the truth condition were instructed to work on the problems alone. All participants were told that after completing these problems, they would be given 5 min to plan what they might say when the experimenter questioned them, but (in order to maintain procedural consistency between this and the RT condition) they were not given any information regarding the specific questions that would be asked. They were informed that after 5 min, the experimenter, who did not actually know whether or not they had cheated, would return to the room and ask them a series of questions to which they should respond by claiming not to have cheated. To motivate participants to tell convincing stories, they were given $15, and were told that if they were able to convince the experimenter that they did not cheat, they would be permitted to keep the money. This manipulation was framed as a loss rather than a gain in order to match the motivation of RT participants who feared the loss of their integrity if it they were judged to have cheated. Next, the research assistant and participant completed their assigned task. After 5 min of planning time, the research assistant signaled for the experimenter to return to the room. The experimenter then accused the students of cheating on the problems and delivered the threat utilized in the RT condition (i.e., having to go before the university disciplinary board), and then asked participants the series of questions to which they responded, as per the research assistant’s instruction, by denying having cheated. Participants were told they could keep the $15, and were debriefed and dismissed. This procedure resulted in a 2 (RT v. MT)  2 (truth v. lie) between-subjects design. All but two participants – both in the RT lie condition – consented to allow their videos to be used in further research and analyses. Note that a high degree of procedural consistency was maintained between the RT and MT methods. Participants in both groups completed the same problems, received the same threat, had the same amount of planning time (5 min) after the delivery of the threat, during which they all expected to have to answer questions but did not know specifically what they would be asked. Furthermore, the positioning of each person in the study (as well as the camera’s angle and distance from the participant) remained consistent across RT and MT conditions. Hence, the only difference between the RT and MT videos was whether the speakers chose to cheat and lie or were asked to do so.

Please cite this article in press as: Hatz, J. L., & Bourgeois, M. J. Anger as a cue to truthfulness. Journal of Experimental Social Psychology (2010), doi:10.1016/ j.jesp.2010.02.009

ARTICLE IN PRESS J.L. Hatz, M.J. Bourgeois / Journal of Experimental Social Psychology xxx (2010) xxx–xxx

3

During Phase 2, participants viewed a total of eight video clips of students telling lies and truths under these two conditions (two from each condition of the 2  2 design), and rated how angry they perceived the speakers to be on a scale of 1–7. Results Planned comparisons revealed that truth-tellers (M = 3.19, sd = 1.019) appeared more angry than lie-tellers (M = 2.54, sd = .959) in the real transgression condition (t(78) = 6.85, p < .001) but not in the mock transgression condition (M truth = 1.91, sd = .767; M lie = 1.93, sd = .872; t(78) = .11, p = .92) (see Fig. 1). Verbal cues

Fig. 2. RT truth-tellers use more anger words than RT liars; again, no difference exists in MT condition.

Methods Transcripts from 154 videos collected via the RT and MT methods (including the 60 videos used in the judging of nonverbal cues described above) were analyzed with Pennebaker’s Linguistic Inquiry & Word Count (LIWC) program (Pennebaker, Booth & Francis, 2001) to assess the percent of anger-related words uttered by participants during the interrogation described above. This program has been used in previous research to compare the speech content of people telling the truth and lying (Newman, Pennebaker, Berry, & Richards, 2003). Results Mean comparisons showed that truth-tellers (M = .732, sd = 1.27) used more anger words than lie-tellers (M = .238, sd = 0.68) in the RT condition (t(59) = 1.90; p = .06) but not in the MT condition (M truth = .815, sd = 1.50; M lie = .806, sd = 0.97; t(26) = .02; p = .986) (see Fig. 2). Discussion Prior research has demonstrated that negative emotion is a reliable cue to deception (Depaulo et al., 2003). Here we have shown support for Ekman’s suggestion that truth-tellers actually express more anger than liars, both verbally and nonverbally; but only in situations where they actually believe themselves to be in trouble (i.e., real transgressions). This finding is notable in that it is the first to show that negativity is not necessarily a cue to deception, and that in some cases it can actually be a cue to truthfulness. Therefore, our results suggest that it may be risky to generalize from the results of the Depaulo et al. (2003) meta-analysis, and assume that any negative emotion signals deception.

Our results also have implications for the study of deception more generally because they underscore the importance of taking into account the specific situational aspects of each lie-eliciting situation. Prior researchers have increased participant motivation by offering monetary rewards for convincing a questioner of their innocence. While this monetary incentive is effective in increasing the intensity of negativity cues (DePaulo et al., 2003), our findings demonstrate that motivations to protect one’s identity or positive self-concept can change not only the intensity, but the quality of deceptive cues. Additionally the present findings lend important insights to the interpretation of polygraph results. It has been well documented that the polygraph is an imprecise measure of deception because it measures physiological arousal, which can accompany any number of emotions (Lykken, 1984). For instance, a suspect taking part in a polygraph examination may be aroused because they are lying and nervous about getting caught, or because they are telling the truth and are nervous or fearful about the situation. We have demonstrated another possibility: that an innocent suspect could fail a polygraph exam due to physiological arousal stemming not from fear or nervousness, but from anger, which also leads to heightened arousal as measured by skin conductance and heart rate (Levenson, Ekman, & Friesen, 1990; Levenson, Ekman, Heider, & Friesen, 1992). Hence, our results support the already widespread observation that polygraph data is not a reliable indicator of deception. Finally, we would like to note the different patterns obtained in our analyses of nonverbal and verbal cues. When judgments are based on nonverbal cues, those in the real transgression condition appeared significantly angrier than those in the mock transgression condition. Analyses based on verbal cues showed the opposite pattern, with those in the mock paid condition using more anger words than those in the ‘real transgression’ condition. This discrepancy supports Ekman’s assertion that nonverbal behaviors – particularly negative facial expressions – are more difficult to feign than the spoken cues. References

Fig. 1. RT truth-tellers appear angrier than RT liars; no difference exists in MT condition.

DePaulo, B. M., Lindsay, J. J., Malone, B. E., Muhlenbruck, L., Charlton, K., & Cooper, H. (2003). Cues to deception. Psychological Bulletin, 129, 74–118. Ekman, P. (1985). Telling lies. New York: Norton. Frank, M. G., & Ekman, P. (1997). The ability to detect deceit generalizes across different types of high-stake lies. Journal of Personality and Social Psychology, 72, 1429–1439. Frank, M. G., & Ekman, P. (2004). Appearing truthful generalizes across different deception situations. Journal of Personality and Social Psychology, 86, 486–495. Iacono, W. G. (2008). Effective policing: Understanding how polygraph tests work and are used. Criminal Justice and Behavior, 35, 1295–1308. Levenson, R. W., Ekman, P., & Friesen, W. V. (1990). Voluntary facial action generates emotion-specific autonomic nervous system activity. Psychophysiology, 27, 363–384.

Please cite this article in press as: Hatz, J. L., & Bourgeois, M. J. Anger as a cue to truthfulness. Journal of Experimental Social Psychology (2010), doi:10.1016/ j.jesp.2010.02.009

ARTICLE IN PRESS 4

J.L. Hatz, M.J. Bourgeois / Journal of Experimental Social Psychology xxx (2010) xxx–xxx

Levenson, R. W., Ekman, P., Heider, K., & Friesen, W. V. (1992). Emotion and autonomic nervous system activity in the Minangkabau of West Sumatra. Journal of Personality and Social Psychology, 62, 972–988. Lykken, D. T. (1984). Trial by polygraph. Behavioral Sciences and The Law, 21, 75–92. Newman, M. L., Pennebaker, J. W., Berry, D. S., & Richards, J. M. (2003). Lying words: Predicting deception from linguistic styles. Personality and Social Psychology Bulletin, 29, 665–675. Pennebaker, J. W., Booth, R. E., & Francis, M. E. (2007). Linguistic inquiry and word count: LIWC2007 – Operator’s manual. Austin, TX: LIWC.net.

Russano, M. B., Meissner, C. A., Narchet, F. M., & Kassin, S. (2005). Investigating true and false confessions within a novel experimental paradigm. Psychological Science, 16, 481–486. Vrij, A. (1995). Behavioral correlates of deception in a simulated police interview. Journal of Psychology: Interdisciplinary and Applied, 129, 15–28. Vrij, A. (2000). Detecting lies and deceit: The psychology of lying and the implications for professional practice. Chichester, England: Wiley. Vrij, A. (2004). Guidelines to catch a liar. In P. A. Granag & L. A. Stromwall (Eds.), The detection of deception in forensic contexts (pp. 287–413). Cambridge University Press.

Please cite this article in press as: Hatz, J. L., & Bourgeois, M. J. Anger as a cue to truthfulness. Journal of Experimental Social Psychology (2010), doi:10.1016/ j.jesp.2010.02.009