Part 1 of 2
Mad and Misleading: Incidental Anger Promotes Deceptionby Jeremy A. Yip and Maurice E. Schweitzer
The Wharton School
8/12/2014
Highlights:
** incidental anger, anger triggered by an unrelated situation, promotes the use of deception.
** deception is also influenced by a number of psychological factors including perceptions of inequity, ego-depletion, power, and trust
** anger may shift attention away from caring about others
** anger promotes punishment, retaliation, and a tendency to rely on stereotypes
** people who feel angry are less cooperative and less interested in interacting with their counterparts in the future
** Angry people may be particularly less concerned about harming others
** Individuals who feel anger may carry their feelings from one interaction to a separate, unrelated interaction
** The greater concern individuals exhibit for themselves and the lower concern for others, the more deceitful they are likely to be
** when individuals lack empathy, they are less concerned about how their actions impact others. As a result, less empathetic people are more likely to behave unethically because they focus more on the rewards for themselves and pay less attention to the costs for others
** when people feel angry, they become more likely to perceive negative events as under the control of others
** anger energizes individuals and promotes the pursuit of self-interested goals
** anger increases the use of deception, but sadness, another negative-valence emotion does not
** When individuals engage in self-interested acts of deception, they derive gains at the expense of others
** anger diminishes empathy for others and disinhibits selfish deceptive behavior
** when participants received $1 for deceiving the taster, participants in the incidental anger condition were more likely to deceive their counterpart (80%) than were those in the neutral condition (62%)
** when participants received no incentive for deceiving their counterpart, none of the participants in the incidental anger condition or the neutral condition deceived their counterpart
** Angry individuals are less empathetic, and as a result, angry individuals care less than neutral individuals about the harmful consequences of their self-interested actions
** people who feel anger are more likely to deceive their counterparts than are people who feel sadness
** anger promotes a range of behaviors from punishment to less careful thinking
** anger promotes deception because when people feel angry, they become particularly motivated to engage in self-interested behavior
** anger reduces empathy
** within a negotiation; displays of anger can elicit concessions
** Individuals who express anger are likely to elicit anger in counterparts
** expressing anger will increase the likelihood that a counterpart will engage in deception
Jeremy A. Yip
The Wharton School
University of Pennsylvania
3730 Walnut Street, 517 JMHH
Philadelphia, PA 19104
Email:
yip@wharton.upenn.eduMaurice E. Schweitzer
The Wharton School
University of Pennsylvania
3730 Walnut Street, 544 JMHH
Philadelphia, PA 19104
Email:
schweitzer@wharton.upenn.eduAcknowledgementsWe thank Madan Pillutla and two anonymous reviewers for their helpful suggestions and guidance. We are grateful for the financial support and research assistance of the Wharton Behavioral Lab. We appreciate the helpful feedback and advice from Kelly Kiyeon Lee, Emma Levine, and members of the OID Schweitzer Lab.
Highlights• We demonstrate that incidental anger, anger triggered by an unrelated situation, promotes unethical behavior.
• People who feel incidental anger are more likely to deceive their counterparts than people who feel neutral emotion.
• Empathy mediates the relationship between incidental anger and deception.
• Incidental anger increases deception relative to sadness.
• Incentives moderate the relationship between incidental anger and deception.AbstractEmotions influence ethical behavior. Across four studies, we demonstrate that incidental anger, anger triggered by an unrelated situation, promotes the use of deception. In Study 1, participants who felt incidental anger were more likely to deceive their counterpart than those who felt neutral emotion. In Study 2, we demonstrate that empathy mediates the relationship between anger and deception. In Study 3, we contrasted anger with sadness. We find that participants who felt incidental anger were more likely to use deception than were participants who felt another negative-valence emotion. In Study 4, we show that incentives moderate the relationship between anger and deception. Collectively, our work reveals that incidental anger promotes unethical behavior because angry people become less empathetic when pursuing their self-interest.Keywords: Lying; Emotion; Anger; Empathy; Negotiation; Conflict; Self-Interest; AdviceMad and Misleading: Incidental Anger Promotes DeceptionDeception pervades organizational life and represents a significant challenge in domains ranging from negotiations to job interviews to expense reporting. In one study, the Coalition Against Insurance Fraud (2012) found that individuals file nearly $80 billion in fraudulent insurance claims in the United States. Financial incentives explain some deceptive behavior (Tenbrunsel, 1998), but recent research suggests that
deception is also influenced by a number of psychological factors including perceptions of inequity (Gino & Pierce, 2010),
ego-depletion (Gino, Schweitzer, Mead, & Ariely, 2011),
power (Pitesa & Thau, 2013),
and trust (Yip & Schweitzer, 2015).
One psychological factor that may be particularly relevant to the deception decision process is emotion (Gino & Shea, 2012; Gaspar & Schweitzer, 2013; Moran & Schweitzer, 2008; Zhong, 2011). In this work, we consider the potential influence of anger on deception. Prior work has linked anger with a number of thoughts and behaviors that are related to deception (Barry & Oliver, 1996; Olekalns & Smith, 2009). For example,
anger curtails cooperation (Lount, Zhong, Sivanathan, & Murnighan, 2008), and increases the rejection of ultimatum game offers (Pillutla & Murnighan, 1996).
In an investigation of expressed anger, Van Dijk, Van Kleef, Steinel, and Van Beest (2008) found that when a counterpart sent a message that expressed anger instead of happiness, people were more likely to send that counterpart incorrect information about the resources available in an ultimatum game.
Surprisingly, no prior research has directly linked feeling angry with deception. This is a surprising omission, because anger is frequently experienced in the workplace in general (Glomb, 2002; Pearson & Porath, 2005) and in negotiations in particular (Van Kleef, De Dreu, & Manstead, 2004; Yip & Schweinsberg, 2016). In our investigation,
we establish a link between feeling angry and deception.By investigating how anger promotes deception, we substantially develop our understanding of both emotion and ethical decision-making. Across four experiments,
we demonstrate that incidental anger, anger triggered by an unrelated source, promotes deception. We also find that feelings of empathy mediate the relationship between anger and deception. We find that anger reduces empathy, which in turn, increases self-serving deception. We also find that incentives moderate the relationship between anger and deception. Collectively, our studies advance our understanding of anger and the psychology of deception.
DeceptionWe focus our investigation on self-serving deception, lies that advantage the deceiver at the expense of the target (Erat & Gneezy, 2012; Levine & Schweitzer, 2014).
Self-serving lies represent a quintessential form of unethical behavior (Gino, Schweitzer, Mead, & Ariely, 2011; Tenbrunsel & Smith-Crowe, 2008), and a growing literature has identified key factors that influence self-serving deception (e.g., O’Connor & Carnevale, 1997; Schweitzer, DeChurch, & Gibson, 2005).
When telling a self-serving lie, deceivers navigate the tension between pursuing their self-interest and harming others. To do this, individuals weigh the potential costs and benefits for themselves (Lewicki, 1983; Loewenstein, Cain, & Sah, 2011) and their counterparts (Gneezy, 2005). Emotions may influence these calculations (Fulmer & Barry, 2009).
In fact, anger promotes a focus on rewards (Aarts, Ruys, Veling, Renes, de Groot, van Nunen, & Geertjes, 2010).
Within the context of self-serving deception, rewards reflect self-interested behavior, and as a result, anger may shift attention toward self-interest.
Other research suggests that anger may shift attention away from caring about others. For example, anger promotes punishment (Fox & Spector, 1999; Wang et al., 2011),
retaliation (Bushman, 2002),
and a tendency to rely on stereotypes (Bodenhausen, Sheppard, & Kramer, 1994). In negotiations,
people who feel angry are less cooperative and less interested in interacting with their counterparts in the future (Allred, Mallozzi, Matsui, & Raia, 1997).
Angry people may be particularly less concerned about harming others. Taken together, we expect anger to lower empathy, and we expect this shift in focus to promote self-serving deception.Emotion and DeceptionEarly work conceptualized ethical decision-making as a cognitive process (Kohlberg, 1969).
More recent work, however, has begun to consider the role that emotions play in ethical decision-making (Haidt, 2001; Huebner, Dwyer, & Hauser, 2009; Pizarro, 2000; Rozin, Lowery, Imada, & Haidt, 1999).
This work has begun to establish a link between emotions and ethical behavior, but scholars have explicitly called for additional research to explore how emotions influence ethical judgment and behavior (Avramova & Inbar, 2013).
Extant work identifies emotions as a consequence of ethical decision-making (Hutcherson & Gross, 2011; Mullen & Skitka, 2006; Rozin, Lowery, Imada, & Haidt, 1999). For example,
unfair ultimatum game offers heightened activity in brain regions associated with emotion (Sanfey, Rilling, Aronson, Nystrom, & Cohen, 2003)
and violations of community standards of fairness elicit anger (Schweitzer & Gibson, 2008).
Similarly, when people’s moral convictions are threatened, people feel angry (Mullen & Skitka, 2006).
Importantly, these feelings can also influence subsequent judgments (Mullen & Skitka, 2006; Schweitzer & Gibson, 2008).
Although several scholars have postulated that emotions are capable of shifting beliefs and behavior (Avramova & Inbar, 2013; Gaspar & Schweitzer, 2013; Huebner, Dwyer, & Hauser, 2009; Wheatly & Haidt, 2005),
surprisingly few empirical studies have directly examined the effects of emotion on ethical behavior (Avramova & Inbar, 2013).
Much of the existing work has focused on envy, guilt, and shame (see Gaspar & Schweitzer, 2013 and Moore & Gino, 2013 for a review).
For example, envy promotes deception (Gino & Pierce, 2009; Moran & Schweitzer, 2008).
In prior investigations, when individuals envied their counterparts, they were more likely to deceive them than when they did not envy them. Similarly, shame may promote deception by exacerbating malevolent intentions (Tangney, Stuewig, & Mashek, 2007) rather than rectifying an underlying problem (Tangney, 1991).
Anxiety also increases deception because anxiety makes individuals feel threatened (Kouchaki & Desai, 2015).
In contrast to envy and shame, feelings of guilt can curtail deception (Zhong, 2011). Surprisingly, prior work has overlooked the potential link between feeling anger and ethical decision-making.
There is limited empirical evidence demonstrating how emotions determine whether an action is right or wrong. Our investigation fills this gap, and more importantly builds our understanding of how emotions influence ethical judgement and behavior.
AngerAnger is a negative-valence emotion that is typically triggered by another person (Dollard, Doob, Miller, Mowrer, & Sears, 1939; Pillutla & Murnighan, 1996; Smith & Ellsworth, 1985).
When individuals blame another person for an injustice, an unfair outcome, or their inability to reach a desired objective, they often feel anger (Lazarus, 1991; Porath & Erez, 2009). Consistent with this conceptualization of anger, prior work has found that
people feel angry when they receive an unfair offer (Pillutla & Murnighan, 1996),
are interrupted (Mauss, Evers, Wilhelm, & Gross, 2006),
read about immoral verdicts (Mullen & Skitka, 2006),
and experience incivility (Pearson & Porath, 2005; Porath & Erez, 2007).
When individuals direct their anger at the offender who treated them unfairly or blocked their goal, they experience directed anger. This anger can prompt individuals to confront, fight or punish the offender (Bushman, 2002; Rozin, Lowery, Imada, & Haidt, 1999). For example, an employee who is insulted by a co-worker may feel anger toward his or her co-worker, and this anger would inform how the employee interacts with that co-worker.
Directed anger reflects the functional nature of emotion (Damasio, 1994).
Anger triggered by one interaction, however, may influence cognition and behavior in an unrelated interaction (Andrade & Ariely, 2009). For example,
the anger an employee feels after a co-worker’s insult may influence that employee’s interactions with his or her spouse in a completely unrelated setting. This influence of anger is incidental and normatively irrelevant to the decision at hand (Dunn & Schweitzer, 2005).
Individuals who feel anger may carry their feelings from one interaction to a separate, unrelated interaction (Berkowitz, 1989).
With incidental emotions, cognitive appraisals may persist beyond the initial emotion-eliciting event (Lerner & Tiedens, 2006; Yip & Côté, 2013).
Anger can shape the perceptions of subsequent, unrelated situations (Dollard et al., 1939; Lerner & Keltner 2001; Schwarz & Clore, 1983; Wiltermuth & Tiedens, 2011).
The study of incidental anger affords both a conservative and a direct test of the influence of emotion on deception. Unlike incidental anger, directed anger confounds emotion with experience. More specifically, individuals who experience directed anger are likely to be motivated by retribution and not just by the emotional experience. In our investigation, we focus on incidental anger and examine whether incidental anger influences deception.
Incidental Anger Increases Self-Serving DeceptionWe advance the following thesis: incidental anger promotes the use of self-serving deception. The decision to engage in self-serving deception balances concern for oneself (i.e. self-interest) and concern for others (i.e. empathy) (Gneezy, 2005; Levine & Schweitzer, 2014).
The greater concern individuals exhibit for themselves and the lower concern for others, the more deceitful they are likely to be. Conversely, if people exhibit lower concern for themselves and they have higher concern for others, they are more likely to tell the truth. We expect anger to diminish concern for others and disinhibit self-interest, which ultimately promotes self-serving deception.Extant work suggests that
empathy influences unethical behavior. Empathy is the capacity to feel emotional concern about the welfare of another party (Davis, 1983).
Pizarro (2000) theorized that empathy sensitizes people about the distress that another person is experiencing, and that a morally-relevant event may be occurring. We reason that when individuals lack empathy, they are less concerned about how their actions impact others. As a result, less empathetic people are more likely to behave unethically because they focus more on the rewards for themselves and pay less attention to the costs for others.Prior research also supports our proposed link between incidental anger and lower empathy. Frijda (1986) suggested that emotions can direct people’s attention towards others for whom they care and away from others. In related work, Pizarro (2000) found that
individuals feel less empathy towards others for whom they blame a violation. Importantly, it is precisely when individuals blame others for a violation that they are likely to feel angry. In addition, when people feel angry, they become more likely to perceive negative events as under the control of others (Fessler, Pillsworth, & Flamson, 2004). Taken together,
we expect angry people to be more likely to blame others, and we postulate that anger will decrease empathy.This expectation is related to prior research suggesting that
anger facilitates self-interest. For example, there is some evidence that
anger energizes individuals and promotes the pursuit of self-interested goals (Aarts et al., 2010; Carver & Harmon-Jones, 2009).
Anger is functional insofar as it enables individuals to attain goals. However, as individuals pursue one salient goal, they may neglect other goals, such as the goal to engage in ethical behavior (Ordonez, Schweitzer, Galinsky, & Bazerman, 2009; Schweitzer, Ordonez, & Douma 2004).
Drawing on prior work,
we expect angry people to become more self-focused than neutral people (Lerner & Tiedens, 2006). Prior work investigating the relationship between feeling angry and punishing others has focused on the motive to harm (Lazarus, 1991; Lerner & Tiedens, 2006; Van Dijk, Van Kleef, Steinel, & Van Beest, 2008). We suggest a distinct, but related account for why angry people punish -- and possibly deceive -- others.
We expect angry people to care less about the harmful consequences of their deception for others and to care more about the beneficial effects of deception for themselves.To summarize,
when individuals encounter an opportunity to engage in self-serving deception, they confront an opportunity to benefit themselves at the expense of others. We propose that as anger reduces empathy, the calculus of this deception decision process changes in a way that promotes deception. More generally, we expect angry people to become more likely to behave unethically at the expense of others as they pursue their self-interest.Overview of Present ResearchWe conducted four laboratory studies to explore the relationship between incidental anger and deception. In Study 1, we induced incidental anger with an essay feedback task and we demonstrate that
incidental anger promotes deception. In Study 2, we use a different incidental anger induction, a recall task, and we replicate our findings in Study 1. In this study, we also identify empathy as a mediator. In Study 3, we compare the influence of incidental anger on deception with incidental sadness and neutral emotion. We find that
anger increases the use of deception, but sadness, another negative-valence emotion does not. In Study 4, we identify incentives as a moderator of the relationship between anger and deception. Across all four studies, we find that incidental anger promotes self-interested deception – even when the target of deception is unrelated to why they are angry. Compared to when people feel neutral emotion, when people feel angry, they are more likely to engage in deception to pursue self-interested goals, because they care less about how their actions affect others.
Study 1In Study 1, we tested our thesis that incidental anger promotes self-interested deception.
Method
ParticipantsWe recruited 230 people from a North American university to participate in a behavioral laboratory experiment for a $10 show-up fee. We randomly assigned half of the people to be participants and half to be evaluators that helped us induce anger. Of the 115 participants, 7 participants failed the comprehension check twice and were not allowed to complete the experiment. No participants reported being suspicious during the study. The final sample size of participants was 108 (M age =21 years, SD age =2.60 years; 71% female).
ProcedureWe began the study with a writing task that both participants and evaluators completed. To induce emotion, we manipulated the feedback participants received. We randomly assigned participants to one of two conditions: Incidental Anger or Neutral. After the emotion induction, participants played a modified version of the deception game.
Emotion Induction. We induced either anger or a neutral emotion by providing participants with feedback on an essay they wrote. Within each experimental session, half of the people were participants and half were evaluators. The evaluators read and evaluated essays, but did not complete any other tasks in our study. That is, the 115 evaluators we recruited provided feedback to create a credible anger manipulation, but they did not provide data for our experiment.
At the start of the session, participants and evaluators wrote an essay by hand for five minutes about an inspirational moment in their lives. Once five minutes had passed, an experimenter collected and exchanged all of the essays. We then asked everyone to read and provide handwritten feedback for another person’s essay.
Unbeknownst to the participants, we gave the evaluators special instructions. Each evaluator provided feedback for a single participant. We asked half of the evaluators to provide negative feedback that included a critical summary of the essay and identified specific problems with the essay (e.g., “This essay describes the following events _____________, which I found to be boring/ordinary/stupid”). We asked the other half of the evaluators to provide neutral feedback and include a factual summary of the essay (e.g., “This essay describes the following events _____________.”). We gave participants and evaluators five minutes to provide feedback. We include the complete set of instructions we gave for writing and evaluating essays in Appendix A. By matching each participant with an evaluator, we were able to provide quick, personalized, and handwritten feedback.
After the feedback portion of the study ended, the experimenter collected the essays with the handwritten feedback and returned each essay to the participant who had authored the essay. Participants had two minutes to read the feedback they received before proceeding to the next task.
We described the next stage of the experiment as a separate study. We administered this stage of the experiment via computer. During this stage of the experiment, we asked Evaluators to complete a filler task. We asked participants to read and make decisions in an Interaction Task.
Interaction Task. In the second half of our study, participants read instructions about a task called the Interaction Task in which they were randomly paired with another participant in the session who did not evaluate their essay. The Interaction Task was a modified version of the Deception Game (Gneezy, 2005; Levine & Schweitzer, 2014; Zhong, 2011). In this task, we assigned participants to the role of the Sender and we paired them with a confederate Receiver.
We presented the Sender with information about two payment options. In OPTION A, the Sender earns $2 and the Receiver earns $1. In OPTION B, the Sender earns $1 and the Receiver earns $2. We told the Sender that the Receiver would know that the two options exist, but would not have information about the payoffs for each option.
The Sender then made a decision to send one of two pre-worded messages to the Receiver: a lie (“OPTION A will earn the Receiver more money than OPTION B”) or the truth (“OPTION B will earn the Receiver more money than OPTION A”). Prior to making the decision about which message to send, we informed the Sender that past research has revealed that the Receiver follows the message sent by the Sender 90% of the time. We also informed the Sender that the identities of the Sender and Receiver would be kept confidential, and that the B money participants earned would be paid to each participant at the end of session privately in cash.
Every participant had to pass a comprehension check. After completing the comprehension check, participants made a decision to send either an honest or deceptive message. Participants then completed an emotion manipulation check and answered demographic questions. We then debriefed and paid participants.
Measures
Anger Manipulation Check.After the interaction task, participants rated the extent to which they felt angry, annoyed, irritated, and outraged (Dunn & Schweitzer, 2005) on a scale ranging from 1 (does not describe my feelings at all) to 7 (describes my feelings very well) (M=2.62, SD=1.93; a=.95).
Deception. We recorded whether participants chose to send a deceptive message (scored as 1) or a truthful message (scored as 0).
Results and DiscussionOur manipulation of incidental anger was successful. Participants in the incidental anger condition reported higher levels of anger (M=3.90, SD=1.90) than did those in the neutral condition (M=1.34, SD=.80), t(106)=-9.15, p<.001.
Supporting our thesis, participants in the incidental anger condition were more likely to deceive their counterparts (82%) than were those in the neutral condition (61%), XX2(1,NN=108)=5.48, p=.019, [x]=.23 (see Figure 1). This finding supports our thesis. Compared to feeling neutral emotion, incidental anger increased deception.
In Study 1, we establish a link between anger and deception. Angry people were more likely to engage in self-serving deception than were neutral people.
Study 2In Study 2, we extend our investigation to explore the role of empathy in mediating the influence of incidental anger on deception. We postulate that anger promotes deception by causing individuals to become less empathetic.
We also extend our investigation by employing a different emotion induction and a different measure of deception. To induce emotion, we used a writing recall task (e.g., Dunn & Schweitzer, 2005; Lerner & Keltner, 2001; Strack, Schwarz, & Gschneidinger, 1985), rather than the essay-feedback task we used in Study 1. To measure deception, we assessed whether or not participants recommend an unpleasant tasting beverage. Participants could earn a bonus payment for misleading their counterpart about the beverage.
Pilot StudyBefore we conducted our main study, however, we conducted a pilot study to link incidental anger with empathy. We conducted our pilot study with an online panel of adult participants on Amazon Mechanical Turk. Our sample consisted of 84 participants (M age=35 years, SD age=12.95 years; 52% female).
As in Study 1, we asked participants in the pilot study to write an essay about an inspirational moment in their lives. We randomly assigned participants to one of two between-participants conditions: Incidental Anger vs. Neutral. Similar to Study 1, participants in the Anger condition received negative, critical feedback. Participants in the Neutral condition received neutral, factual feedback. We provided the same critical feedback to all participants in the Anger condition and the same factual feedback to all participants in the Neutral condition. We include the feedback that we presented to participants in Appendix B.
To measure empathy, we asked participants to complete a measure that we adapted from Davis (1983); M=5.25, SD=1.23, [x]=.80. Our empathy measure is presented in Appendix C. In this study, we explore whether or not anger decreases empathy.
Supporting our thesis, we found that participants in the anger condition reported lower levels of empathy (M=4.85, SD=1.46) than did participants in the neutral condition (M=5.59, SD=.89), t(80)=2.78, p=.007. These findings reveal that compared to people who feel neutral emotion, people who feel angry care less about others.
Method
ParticipantsWe recruited 184 students from a large North American university to participate in this study. We lost data from six participants because of a technical glitch with our instant messaging platform and three participants indicated suspicion. The final sample included 175 students (M age=21 years, SD age=2.40 years; 72% female).
ProcedureAs in Study 1, we informed participants that they would complete separate studies in the same session. We randomly assigned participants to either an incidental anger condition or a neutral condition. We manipulated emotion by having participants complete a writing recall task (Strack, Schwarz, & Gschneidinger, 1985). Participants in the incidental anger condition described a situation that made them the most angry they have ever felt in their lives. Participants in the neutral condition described how they spend a typical evening.
After completing the emotion induction, participants proceeded to a purportedly unrelated
cola beverage taste task. We conducted this part of the experiment via an instant messaging platform. In this task, we assigned participants to the role of advisor and we paired them with a confederate who we assigned to the role of taster. We informed participants that the taster would need to choose one beverage to taste among four beverages labelled A, B, C, or D. We gave participants in the advisor role information about how each beverage tasted. We told advisors that this information would not be available to the tasters. This information took the form of comments from an earlier pilot study.
Three of the beverages (A, C, and D) had comments indicating that they were pleasant tasting (e.g., “Very subtle, smooth.”, “There’s a cherry-ish note, I like that one.”). However, one beverage (Beverage B) was described as having an unpleasant taste (e.g., “That tasted like it had ammonia in it.”).
We provided participants with incentives. Specifically, we told participants that as advisors they would receive $1 in addition to the standard show-up fee if they could persuade the taster to drink beverage B. To enhance credibility, we arranged 16 cups filled with cola organized in four rows on a table near the door of the behavioral lab. Every participant saw the cups filled with cola as they entered the behavioral laboratory. After being seated, we pulled a screen to partition the room in half.
We then asked participants to send messages to their counterpart (the taster) using an instant messaging platform. The instant messaging platform was designed so that the confederate taster would send the first message, the participant advisor would send the second message, and the two parties could continue sending messages. The confederate taster adhered to scripted messages throughout the chat and started the chat by asking the advisor which beverage they would recommend. We include the script for the confederate taster in Appendix D. We assessed deception by recording whether the advisor recommended the unpleasant tasting beverage (Beverage B) or recommended one of the pleasant tasting beverages (A, C, or D).
After the beverage taste task, participants completed an empathy measure followed by an emotion manipulation check and demographic questions. In each emotion condition, we counterbalanced whether the empathy measure was presented before the beverage taste task or after the beverage taste task.
Measures
Anger Manipulation Check. After the beverage taste task, participants rated the extent to which they felt angry, annoyed, irritated, and outraged on a scale ranging from 1 (does not describe my feelings at all) to 7 (describes my feelings very well) (M=3.42, SD=2.39; a=.91).
Empathy. We adapted five items from the interpersonal reactivity index that assessed other-oriented feelings of concern towards unfortunate others (Davis, 1983). Participants rated each item (e.g., “I felt concerned for people less fortunate than me.”) on a scale ranging from 1 (does not describe me at all) to 7 (describes me very well) (M=5.15, SD=1.07; a=.71). We include the complete scale that we used in Appendix C.
Deception. We recorded whether participants deceived their counterpart by recommending beverage B (scored as 1) or were truthful to their counterpart by recommending beverage A, C, or D (scored as 0).
ResultsAs expected,
participants in the incidental anger condition (M=5.25, SD=1.91)
reported higher levels of anger than did those in the neutral condition (M=1.52, SD=.87), t(171)=-16.42, p<.001.
To test our thesis, we conducted a binary logistic regression. We find similar results with a chi-square test for independence because the dependent variable consisted of a binary decision to deceive (recommend beverage B) or tell the truth (recommend beverage A, C, or D).
We found that participants in the incidental anger condition were more likely to deceive their counterpart (82%) than were participants in the neutral condition (52%), [x]=.40, SE=.35, Wald(1)=17.33, p<.001, Φ=.32. Using binary logistic regression analysis, we found no significant effect for order (whether the empathy measure was presented before the beverage task or after the beverage task) on deception, β=.27, SE=.32, Wald(1)=.70, p=.404. We also found no significant effect for the interaction between the emotion condition and order, b=-.79, SE=.72, Wald(1)=1.21, p=.271.
We tested whether empathy mediates the effect of incidental anger on deception by employing the indirect bootstrapping technique (Preacher & Hayes, 2008). As recommended by Zhao, Lynch, and Chen (2010), we performed 10,000 bootstrap resamples using Preacher and Hayes’s (2008) SAS macro. Our analysis revealed that incidental anger had an indirect effect on deception through empathy (b=.18, 95% confidence interval [CI]=.019, .432). Because the bias-corrected 95% confidence interval did not include zero, we conclude that empathy mediates the effect of incidental anger on deception.
In Study 2, we find that when people feel angry, they become more likely to engage in deception and that reduced empathy mediates this relationship.
Study 3In Study 3, we extend our investigation of the link between anger and deception by contrasting anger with another negative-valence emotion, sadness.
Sadness is a negative-valence emotion associated with greater risk-taking (Raghunathan & Pham, 1999),
a greater focus on the present (Lerner, Li, & Weber, 2012),
and more systematic thinking (Tiedens & Linton, 2001). Prior studies have found that similarly valenced emotions, such as fear, anger, and sadness, influence attitudes and behavior very differently. That is, even though emotions are similar in valence, they are distinct from each other along other dimensions or “appraisal tendencies” (Raghunathan & Pham, 1999; Wiltermuth & Tiedens, 2011). For example,
when people feel angry, they become more likely to blame someone else (other-person control). In contrast, when people feel sad, they become more likely to blame the situation (situational control) (Keltner, Ellsworth, & Edwards, 1993; Smith & Ellsworth, 1985).
We do not expect negative-valence to promote deception. Rather, we expect anger to uniquely promote deception, because
anger is likely to decrease concern for others. In contrast, sadness may cause people to engage in systematic thinking (Tiedens & Linton, 2001)
and remain concerned about the potential harm that their actions may cause others. Taken together, we expect that incidental anger promotes self-serving deception compared to incidental sadness and neutral emotion.
Method
ParticipantsWe recruited 190 students at a large North American university for a $10 show-up fee. Eleven participants failed the comprehension check twice for the interaction task and, therefore, did not complete the study. Our final sample consisted of 179 participants (M age=22 years, SD age=4.92 years; 57% female).
ProcedureWe randomly assigned participants to one of three conditions: Incidental Anger, Incidental Sadness, or Neutral Emotion.
To manipulate emotions, we had participants watch emotion-inducing video clips (Gross & Levenson, 1995). Specifically,
participants in the incidental anger condition watched a video clip of a woman yelling racial epithets at a convenience store clerk. Participants in the incidental sadness condition watched a video clip from the cartoon movie, Up; in this clip, participants watch a husband and wife grow old and see the husband’s wife pass away. Participants in the neutral condition watched a video clip from a National Geographic documentary about ocean life. All video clips are available upon request from the authors.
After watching the video, we asked participants to describe the video in one sentence, rate the resolution quality and rate the sound quality of the video. These questions misdirect participants so that they could be less likely to attribute their feelings to the video clips.
Participants then continued to a purportedly unrelated study, which we called the Interaction Task. We used the same paradigm to measure deception as we did in Study 1. However, we presented different payment options to the Sender than we did in Study 1. In OPTION A, the Sender earns $1.50 and the Receiver earns $1.25. In OPTION B, the Sender earns $1.25 and the Receiver earns $1.50. We assessed whether participants sent a deceitful message or truthful message to their confederate counterpart.
Following the Interaction Task, participants completed an emotion manipulation check and answered some demographic questions. We then debriefed and paid participants.
Measures
Emotion Manipulation Check. After the Interaction Task, participants rated the extent to which they felt different emotions on a scale ranging from 1 (does not describe my feelings at all) to 7 (describes my feelings very well). We measured anger by averaging responses for angry, irritated, and annoyed (M=2.72, SD=2.08; a=.97). We measured sadness by averaging responses for sad, down, and gloomy (M=3.11, SD=1.73; a=.90). We measured neutral emotion by averaging responses for indifferent, neutral, and calm (M=3.12, SD=1.55; a=.74).
Deception. We assessed whether participants chose to send a deceitful message (scored as 1) or a truthful message (scored as 0).
Results and DiscussionOur emotion induction was effective. We found that
participants reported higher levels of anger in the incidental anger condition (M=4.97, SD=1.79)
than participants in the sadness condition (M=1.21, SD=.53)
or the neutral condition (M=1.90, SD=1.17), F(2, 176)=148.44, p<.001. We report results for the emotion manipulation checks in Table 1.
As predicted,
participants in the incidental anger condition were more likely to deceive others (77%) than were those in the sadness condition (54%) and the neutral condition (61%), XX2(2,NN=179)=7.19, p=.027, [x]=.20.
We conducted separate chi-square tests to compare the influence of emotions on deception. As found in our previous studies, participants in the anger condition were more likely to engage in deception than were participants in the neutral condition, XX2(1,NN=120)=3.65, p=.057, Φ=.17. We also find that participants in the anger condition were more likely to deceive others than were those in the sadness condition, XX2(1,NN=120)=6.94, p=.008, [x]=.24. There was no difference in deception between participants in the neutral condition and sadness condition, XX2(1,NN=118)=.56, p=.456, [x]=.07.
We find that negative-valence alone does not have an effect on unethical behavior. We demonstrate that incidental anger has a unique influence on deception that is separate from another negative-valence emotion, sadness.