Altruistic punishment vs. altruistic compensation in
cases of injustice: What is the role of empathy?
Research Master Thesis submitted by
Manolis Michelakakis
s2042657
1
stSupervisor: dr. A.A.J. (André) van Hoorn
2
ndSupervisor: dr Marijke C. Leliveld
Faculty of Economics and Business
University of Groningen
Abstract
When justice is violated, observers can react altruistically, by punishing the violator or by compensating the victim. Although altruistic behavior is strongly determined by empathy, previous studies have ignored to study its role in third party reactions to injustice. Moreover, the vast majority of previous studies have focused on altruistic punishment and overlooked the importance of altruistic compensation. In this study we explored third parties reactions to injustice. Specifically, we studied preferences for altruistic punishment versus altruistic compensation and we experimentally tested whether preferences are affected when manipulating the empathic concern towards the injustice victim. Overall, altruistic compensation was found to play an important role for restoring justice since it was chosen as an option from the majority of our sample. Participants in both low-empathy and high-empathy conditions perceived the injustice similarly. Yet, the former showed a strong preference for punishment of the justice violator, whereas in the latter the levels of compensation for the victim and of punishment for the violator did not differ significantly. These findings provide useful insights for real life situations when observers’ reaction to injustice is targeted, such as in humanitarian campaigns.
3
1. Introduction
1.1 Altruistic Punishment as norm enforcement
Human societies are characterized widely by social norms, or else behavioral standards and normative group-held beliefs about how individual group members ought to behave in a given context (Hechter & Opp, 2001). Social norms influence the vast majority of human lives’ aspects and relationships such as justice, cooperation, politeness, trust and competition. Yet, little is known concerning the cognitive and emotional basis of social norms, their influence on our attitudes and reactions as well as the motivations behind their enforcement (Fehr & Fishbacher, 2004). The latter is particularly important as it is highly unlikely for everyone in a specific society or a social group to share exactly the same norms based on the same individual and group values. Therefore, certain types of informal or formal social sanctions and rewards are required in order to secure norm compliance (Fehr et al., 2002; Fehr & Fishbacher, 2004). One of the most prominent means of norm enforcement, altruistic punishment, has during the last decade attracted the attention of different research fields such as psychology, biology, anthropology, philosophy and especially economics (Fehr & Gächter, 2002).
Altruistic or else costly punishment refers to punishment imposed to a norm violator by an agent at his/her personal cost even when any direct or future payoff to this agent is unlikely. This type of punishment to norm violations is a fundamental characteristic of human altruism and is closely related to the concept of strong reciprocity or else the reciprocity between non-related persons or persons that will not interact again in the future and their reputation is not depended their interaction (Fehr & Gächter, 2002; Fehr et al., 2002; Gintis et al., 2003; Fehr & Fishbacher, 2003;). Altruistic punishment and its effectiveness for enforcement of norms such as cooperation and justice is a widely established phenomenon not only in industrialized societies but also across diverse populations from around the globe (Henrich et al., 2005). Interestingly, this concept of altruistic punishment is in direct contrast with the self-interest and rationality assumptions for human behavior which are the core of economics canonical model (Henrich et al., 2006).
1.2. Altruistic punishment from third parties
Altruistic punishment may originate either from a second party which was directly hurt by the norm violation or by an unaffected third party. The existence of sanctions from third parties is very crucial for compliance in social norms since the victims of norm violators can be restricted only within a small or powerless group of second parties. Therefore punishment only from the victims would have very limited scope for norm enforcement in regulating social life (Fehr & Fishbacher, 2004).
himself/herself and the recipient who has no choice but to accept it. In the third party punishment dictator game there is a third party, who observes the distribution and then decides to either punish the dictator or not. In Fehr & Fishbacher (2004)’s study, 55% of the third parties punished the violation of the equal distribution norm and the punishment was increased for stronger norm violations. This is particularly interesting, given that punishment is costly and hence no self-interested third party would decide to punish unless a distributional justice norm does exist. This study therefore demonstrates the relevance of altruistic punishment in the case of directly unaffected third parties.
1.2. Altruistic compensation from third parties
As we explained, altruistic punishment serves as a means of enforcement of norms such as justice and deterrence of future violation. However, when people observe injustice and have the motivation to react to it, they might choose to punish or to compensate or both, although punishment is more common (Darley & Pittman, 2003; Gromet & Darley, 2009; van Prooijen, 2010). Despite the fact that both punishment and compensation aim to restore justice and equality, their motivations and focus are different. While the former focuses on the violator and aims at retribution and future deterrence, the latter is motivated by the desire to restore the victim’s condition (Gromet & Darley, 2009). Despite compensation’s relevance, the vast majority of the literature in third party reactions to justice violation has focused on punishment exclusively without having both options available (Lotz et al., 2011; Leliveld et al., 2012).
1.3. Empathy, altruism and economics
The importance of third party altruistic punishment in cases of justice violations is already well established as we explained. Yet, there is little research on the motivations behind third party punishment. Especially moral emotions such as guilt and empathy, which have been recognized by psychologists as very prominent motivations, have been largely ignored by economists who have traditionally focused on the aspects of rationality and self-interest (Fehr & Gachter, 2002; Gintis et al., 2003; Fehr & Fishbacher, 2004; Bowles & Gintis, 2006; Frith & Singer 2008; Bowles, 2008).
Empathy in particular, described as the ability to understand and share the emotions of others, has been argued to be a fundamental basis of altruism and pro-social behavior (Batson, 1991; Davis, 1994; Eisenberg & Miller, 1987; Hoffman, 2000; de Waal, 2008). Moreover, the importance of empathy for economics was firstly suggested at least back in the eighteenth century (Fontaine 1997, 2001) when Adam Smith, the “father” of modern economics wrote his renowned, “The theory of moral sentiments” (1759). Adam Smith suggested that “sympathy” (defined closer to the modern term of empathy) is the basis for moral behaviors and acts as social glue that connects us to others. Quoting from the very first line of his book:
“How selfish soever man may be supposed, there are evidently some principles in his nature,
5
made to conceive it in a very lively manner. That we often derive sorrow from the sorrow of others, is a matter of fact too obvious to require any instances to prove it;”
The more general role of empathy, as social glue in the society, appears prominent also in his Wealth of Nations since it is crucial for a successful trade between economic agents (Fontaine 1997).
However, the concept of empathy did not attract the deserved attention after Adam Smith and especially during the 20th century when the aspects of human rationality, self-interest
and market anonymity played a fundamental role in economic theory. Recently though, especially due to the development of behavioral and neuro economics, economists’ interest in empathy has revived (Singer & Fehr, 2005; Loewenstein & Small 2007; Frith & Singer 2008; Kirman & Teschl, 2010). As Singer & Fehr (2005) argue, the assumption that agents behave strategically because of their ability to understand and predict the reactions and motivations of others is central in game theory. Hence, the ability to understand and share the feelings of others appears very crucial (Singer & Fehr, 2005). In these terms, empathy appears to serve our self-regarding preferences through understanding the reactions and motivations of others. At the same time, it also serves our other-regarding preferences by sharing the others’ feelings and promotes altruistic behavior (Singer & Fehr, 2005).
Moreover, neuroscientific experiments suggest that empathic brain reactions from third parties during economic games are dependent on the perceived fairness of the involved parties (Singer et al., 2006). Singer et al. (2006), in theirs prisoner’s dilemma game experiment, had participants to observe two confederates first playing fairly or unfairly and then receiving pain. Under this setting participants exhibited empathy related brain activity when observing the fair confederate receiving pain; the activity was significantly reduced for the unfair confederate. Moreover, in the case of male participants, knowing that the unfair confederate received pain activated reward-related brain areas. This study’s result that learned preferences modulate empathic responses is quite consistent also with relatively recent economic theories of social preferences (Rabin, 1993; Fehr & Schmidt, 1999).
1.4. Purpose of the study
Still, despite the important role of empathy in altruistic and economic behavior and the equally important role of altruistic punishment in reactions to injustice there is currently only one study which explored their relationship (Leliveld et al. 2012). Altruistic compensation is also similarly ignored from the respective literature, despite its relevance as a third party reaction in cases of injustice (Lotz et al., 2011; Leliveld et al. 2012). This is the important gap that we intended to fill with our study. In particular, we explored how the inducement of empathic concern towards the victim of an injustice affected the reactions of third party observers with respect to altruistic punishment versus altruistic compensation. The key of our research is the fact that we experimentally manipulated empathic concern instead of just measuring it as a general human trait as Leliveld et al. (2012) did.
reactions to injustice. People are daily informed about cases in which an agent (e.g. individuals, group of people, institutions etc.) has committed injustice towards a “victim”. Entities such as individuals, social groups, NGOs and governments aim specifically to communicate this injustice to society members who are expected to take action. Understanding how the inducement of empathic concern towards the victim might affect expected reactions is particularly important when deciding how to communicate the injustice. For instance, if the goal is to intrigue victims’ compensation instead of violators’ punishment, then it might be more effective when the injustice information focuses on the suffering of the victim and trying to communicate the victim’s perspective instead of focusing on the violator or the other way around. Since the amount of action by those who are willing to react upon an observed injustice is limited, different reactions like punishment and compensation might “compete” with each other. A relevant real life example might be the case of a company polluting the environment and having negative externalities to the lives of a specific group of people. A third unaffected party in that case could be a group of this company’s customers which have both the choice to boycott and to donate.
In order to study our research question, we utilized a third party punishment/compensation game. This game is based on Fehr & Fishbacher’s (2004) third party punishment dictator game but has altruistic compensation as an available option too. The third party punishment / compensation game is similar to the one used by Leliveld et al. (2012), but not quite. In our study we also allowed the choice of , both compensating the victim and punishing the offender was also available for the third party since we believe that such a setting is more realistic to real life situations. We also manipulated empathic concern towards the victim in line with the empathy manipulation of Batson et al. (2007).
Previous research (Leliveld et al., 2012), suggests that high and low empathic people did not differ in their perceptions of justice and both were willing to restore it in case of violation. However, high empathic people showed a preference for the altruistic compensation of the victim (recipient) whereas low empathic people preferred to restore justice by altruistically punishing the justice violator (dictator).
Based on these results, we hypothesized that the group of participants who were led to feel empathic concern towards the victim of the unfair distribution would choose to compensate him/her more often and to a greater extent compared to the group of participants who were not primed with empathic concern. On the other hand, participants in the latter group were expected to choose more frequently punishment and at higher levels compared to the ones in the former. However, the influence of empathy manipulation on the preference between compensation and punishment cannot be hypothesized beforehand. The reason is that empathic concern towards a victim promotes helping behavior (Batson, 1991; Davis, 1994; Eisenberg & Miller, 1987; Hoffman, 2000) but on the other hand high empathic people perceive greater reward when they observe the punishment of an unfair player (Singer et al., 2006) compared to the low empathic ones.
7 important role in a game setting like ours (Batson et al., 2007; Nelissen & Zeelenberg, 2009; Lotz et al., 2011).
2. Method 2.1. Participants
Participants in the experiment were 81 students at the University of Groningen and the Hanze University of Applied Sciences Groningen (46 male, mean age 21.38, SD=3.33). They received 2 research credits or 3 Euros as a compensation for participation in the 20 minutes study plus the chance to win one of the three 30 Euros lottery prizes. The experimental session took place in the research lab of The Faculty of Economics and Business during the first two weeks of June in 2013.
2.2. Procedure
At the start of the study, participants were told that they were going to take part in “A Bargaining Interaction” between three parties under complete anonymity. The instructions explained that the study’s aim was to explore how people act in a certain type of bargaining interaction. They were also led to believe that a random draw linked them with two other present participants in different cubicles in the lab and assigned each one of them to one of the three possible roles, Player A, B, or C. In fact, all participants were assigned the role of Player C or else the observer of a one-shot dictator game between Player A (the dictator) and Player B (the recipient). Participants, then, were informed that they were endowed with 50 points while player A was endowed with 100 points. Each point represents a lottery ticket. Player A in phase 1 had to decide among six alternatives namely to transfer 0, 10, 20, 30, 40, or 50 to Player B, who had no endowment (see Fehr and Fischbacher, 2004). Player B had to accept Player A’s distribution and could not influence it. In phase 2, they, as observers of the distribution chosen by Player A, had the opportunity to influence the final distribution by punishing Player A, compensating Player B or a combination of the previous two through transferring part of their points. Every point they would transfer to Player B would increase his/her points by three points and respectively every point transferred to Player A would reduce his/her endowment by three points. In order to check whether the participants understood correctly the rules of the interaction, we also presented to them a hypothetical example and asked them to give the correct answer concerning the final outcome.
mentioned in the instructions neither the terms punish, compensate. Instead, we used Player A, B, C and reduce/increase respectively.
2.3. Empathy manipulation
Previous research on empathy (Coke et al., 1978; Batson, 1991) suggests that in order to induce empathy, two conditions must be met: perception of another person as in need and adoption of that other's perspective. For that reason we introduced three experimental conditions based on Batson et al. (2007): low-empathy (LE), high-empathy (HE) and no-communication (NC). We informed the participants in the two no-communication conditions (HE and LE) that during the interaction there would be a simple form of communication. For this purpose, one of the three players would be randomly assigned to be the Sender of communication. The Sender would write a brief note describing something of interest that occurred in his or her life recently. A second player would be randomly assigned to be the Receiver of communication. The Sender was supposed to have written the note before receiving the instructions for the bargaining interaction so that he/she would not be influenced by such knowledge. Respectively, all participants in those two conditions learned that they were assigned (ostensibly randomly) to the role of the Receiver of the note before knowing their role as Player C. The fake note revealed the need of Player B for cheering up because of a recent romantic break-up. However, in the HE condition the participants were instructed to read the note from its writer’s perspective imagining how he/she felt while in the LE condition they were instructed to read it from an objective perspective. Through this manipulation we make sure that any observed effect is a product of empathy and not simply of receiving communication or knowing that the other player is in need. Finally, in order to minimize cross-gender self-presentation concerns (Jones & Pittman, 1982) we presented the personal note as being written either from a male or a female depending on the gender of the participant. The manipulation that we applied has successfully induced empathic concern in past research through stimulating emotions and physiological arousal, (Batson, 1991). In the third condition (NC) there was no information concerning communication neither a note. Through the use of this last group we aimed to separate further the effects of communication.
2.4. Measures
Empathy was was mesured after participants read the note, by using a Likert-type self-reported feelings questionnaire. In this questionnaire, participants had to indicate their current feelings towards Player B through a list of 14 emotion adjectives (1=not at all, 7=extremely). Six of the items included were empathetic oriented adjectives (see Batson, 1991): sympathetic, warm, compassionate, softhearted, tender, and moved. The responses from the six empathetic oriented items were averaged to an empathy index (Cronbach’s
α=.86). Previous research has found that those 6 adjectives indeed form a single factor that
measures empathy distinct both from distress and general emotionality with a Cronbach’s α ranging from .91 to .93 (Batson, 1991; Batson et al., 2007).
9 “appropriate” was the offer (1= not at all, 7=extremely) (Leliveld et al., 2012). These questions were posed in the end of the study and were combined into a justice scale (α=.89).
Moreover, we included a number of scales and questions (all measured in a scale from 1 to 7) in order to explore further the emotions and motivations for the observed participants’ choices. Firstly, we incorporated a scale for anger which is suggested as a very common feeling behind third parties’ reactions when confronting injustice (Batson et al., 2007; Nelissen & Zeelenberg, 2009; Lotz et al., 2011). Anger was measured by reporting to what extent the participant felt, among some other filler emotions, “anger”, “mad”, “annoyed”,
“offended”, “outraged”, “irritated” towards Player A (Cronbach’s α=.95). Secondly, we
measured feelings of guilt and the related sense of responsibility based on the findings of Nelissen & Zeelenberg (2009) and Lotz et al. (2011). Particpants’ guilt was measured by indicating whether they felt “guilt”, “bad”, “ashamed” for not assigning increase points to Player B (guiltB, Cronbach’s α=.89) or decrease points to Player A (guiltA, Cronbach’s α=.85) . In order to check whether the reaction of participants was influenced by the extent they felt responsible for taking action towards an unequal distribution we formed a responsibility index (Cronbach’s α=.89) in which they indicated whether they felt “responsible”, “accountable” and “liable” for assigning points to Player B and/or Player A. Finally, we included a few more questions in order to explore further participants’ motivations (see Table 1).
Table 1. ‘Motivation’ questions
Question Variable
Did you intend to compensate Player B because of Player A’s decision? compensate Did you intend to punish Player A for his/her decision? punish
Did you intend to make the outcome of Player B more equal to the one of Player A? equal
Did you intend to make clear to Player B that you wanted to help him/her? help
Did you intend to make clear to Player A that he/she violated a social norm? violated
Did you intend to maximize your income? maximize
To what extent did you find it important not to influence Player A’s and B’s number of points? influence
To what extent do you think Player A is friendly? friendly
To what extent do you think Player A is selfish? selfish
If you would have been Player A, which offer (0, 10, 20, 30, 40, 50) would you have made to Player B? offer
motivation behind assigning increase points to Player A and/or reduction points to Player B and the last one was offered as a probe for suspicion.
3. Results
3.1. Reactions of participants in the whole sample
In total 19.8% of the participants chose to keep all their points, 12.3% only to compensate, 27.2% opted only to punish and 40.7% chose both, χ2(3)=14.259, p<.01 (see Table 2). Hence,
we see that the choice of assigning both punishment points to Player A and compensation points to Player B was the most preferred. Overall, 17.3% of the total 81 participants chose to transfer more compensation points than punishment points, 49.4% chose more punishment points and 33.3% transferred exactly the same amount of points for punishment and compensation, χ2(2)=12.519, p<.01. Hence, almost half of the participants showed a
preference for punishment whereas one third of the participants did not express a preference for one of the two strategies.
Concerning the amount of points, on average 4.26 (SD=8.45) points were transferred as compensation and 9.15 (SD=8.53) as punishment, with a Wilcoxon signedrank test of z = -3.99, p <.001, which makes an average total of 13.41 (SD=10.61) transferred points. In order to derive clearer conclusions from the comparison between compensation and punishment since most of the participants chose for both, we split all the participants into three distinct and mutually exclusive groups: the ones who assigned more compensation than punishment; the ones who assigned more punishment than compensation; and the ones who assigned equal punishment and compensation. Based on the Kruskal-Wallis tests, all these three groups were significantly different in the mean compensation points H(2)=22.8,
p<.001, the mean punishment points H(2)=55.78, p<.001 and the total assigned points H(2)=28.76, p<.001. Moreover, we observe that participants who chose to transfer more
compensation points than punishment points assigned on average 14.93 (SD=16.23) total points whereas the ones who chose punishment as a dominant strategy transferred on average 18.13 (SD=5.94) points, with a Mann-Whitney test of U=150.5, p<.01. Hence, the preference for punishment results on an average of 3.2 more total assigned points than the preference for compensation.
11
Table 2. Reaction of participants
Reaction strategy of participants
Percent of participants
Total points Compensation points
Punishment points M (SD) M (SD) M (SD) Keep all points: 19.8 - - - Compensate: 12.3 15.50(18.35) 15.50 (18.35) - Punish: 27.2 18.23(3.46) - 18.23 (3.46) Both: Comp>Punish: 4.9 13.5(11.27) 10.75 (9.95) 2.75 (1.71) Punish>Comp: 22.2 18(8.14) 3.94 (2.16) 14.06 (6.94) Equal: 13.6 13.82(6.35) 6.91 (3.18) 6.91 (3.18) Total: 40.7 16.06 (8.02) 5.76 (4.51) 10.30 (6.93) Compensate>Punish: 17.3 14.93 (16.23) 14.14 (16.15) 0.79 (1.52) Punish> Compensate: 49.4 18.13 (5.94) 1.78 (2.44) 16.35 (5.65) Equal: 33.3 5.63 (7.96) 2.81 (3.98) 2.81 (3.98) All: 100 13.41 (10.61) 4.26 (8.45) 9.15 (8.53)
3.2. Empathy manipulation check
As expected, participants who read the note according to the HE condition scored higher in the self-reported empathy index (M = 4.05, SD = 1.25) compared to participants in the LE condition (M = 3.66, SD = 1.26). Nevertheless, this difference was not statistically significant
t(54)=-1.17, p=.12 (one tailed) and its effect was less than medium, Cohen’s d=.31.
Therefore, we cannot suggest that the manipulation was effective as measured by our specific empathy scale.
However, comparing the empathy scores of every subgroup between the two conditions (LE vs. HE) we observe that empathy is always lower in the LE condition but again none of the differences is statistically significant. The lowest empathy scores were observed in the LE condition and specifically in the subgroup that chose to keep all the points (M=3.03, SD=1.59) and the subgroup that chose only punishment (M=3.5, SD=1.29) as hypothesized. On the other hand, the highest empathy scores were again as hypothesized in the HE condition and specifically the subgroup that chose only to compensate (M=4.39, SD=1.77) and the subgroup that chose both punishment and compensation (M=4.12, SD=1.12). The difference in empathy that was closer to statistical significance was between the LE - keep all group and HE- both, t(16)=-1.63, p=0,059 (one tailed).
Turning to the comparison between HE and LE condition we evidence that according to our hypothesis there was a greater percentage of participants in the HE condition who showed a preference for compensation (21.7%) compared to the LE condition (10.7%) and at the same time a smaller preference for punishment (39.3%) than in LE (53.6%) (see Table 3). However, those differences appear statistically insignificant and cannot be compared accurately with chi-square test since the minimum expected cells count is not met. On the other hand, a binomial comparison of the two percentages within each condition indicated that although the difference within the LE condition (10.7% vs 53.6%) was statistically significant, p=.008, the respective one for the HE (21.4% vs. 39.3%) was not, p=.33.
Concerning the actual amount of points transferred, although the difference again is towards our hypothesis, or else participants in HE transferred on average more compensation points (M=5.46, SD=10.29) compared to LE (M=2.46, SD=3.26), it was not statistically significant, U=335, p=.16 (one tailed) and its effect size was between small and medium with a Cohen’s d=0.4. Similarly, HE conditioned participants assigned less punishment points (M=7.82, SD=8.87) compared to LE condition (M=9.54, SD=8.50) but again this difference is even less significant than the previous one. Overall, the mean total points assigned from the participants in the two conditions (M=12, SD=8.45 vs. M=13.29, SD=12.3) did not differ significantly, U=378, p= .817. However, there is a significant trend towards our hypothesis on the difference between the compensation points and punishment points within each condition which appears to be quite smaller and statistically insignificant for the HE (M=5.46, SD=10.29 vs. M=7.82, SD=8.87), z = -1.47, p =.141, compared to the larger and statistically significant one within the LE (M=2.46, SD=3.26 vs. M=9.54, SD=8.5), z = -3.15, p=.002. Moreover, the compensation points transferred from the specific subgroup of participants who showed a preference for compensation within the group of HE condition (M=18, SD=17.27) were significantly higher compared to the same amount within the LE condition (M=6, SD=3.61), U=2.5, p= .043 (one-tailed) with a very large effect size, Cohen’s d=.96.
With respect to the final distribution of points, participants in the HE condition chose on average 56.54 points for Player A, 36.38 points for Player B and 36.71 for themselves. Participants conditioned with LE chose a mean final distribution with similar equality namely, 51.38 points for Player A, 27.38 points for Player B and 38 for themselves. Those two distributions do not differ significantly either on the gap between Player A’s and Player B’s points or the sum of the gaps between each player’s points and 50. This intention is also confirmed by the fact that the two groups scored on average very similarly in the “equal” variable (see Table 1.), namely 3.82 (SD=2.28) for the LE and 3.46 (SD=2.5) for the HE condition, t(54)=.56, p= .58. Similarly, they did not score significantly different neither in the justice scale, with a mean score of 3.87 (SD=1.03) for the LE and 3.62 (SD=1.64) for the HE,
13
Table 3. Reaction of participants in each condition
Condition
Reaction strategy of participants
Percent of
participants Total points
Compensation points Punishment points M (SD) M (SD) M (SD) Low-empathy (28) Compens>Punish 10.7 6 (3.61) 6 (3.61) - Punish> Compens 53.6 17.07 (5.68) 1.33 (1.78) 15.73 (6.16) Equal 35.7 6.2 (8.3) 3.1 (4.15) 3.1 (4.15) All 100 12 (8.45) 2.46 (3.26) 9.54 (8.50) High-empathy (28) Compens>Punish 21.4 19.33 (17.34) 18 (17.27) 1.33 (1.97) Punish> Compens 39.3 19.63 (6.65) 2.27 (2.33) 17.36 (5.85) Equal 39.3 3.64 (6.74) 1.82 (3.37) 1.82 (3.37) All 100 13.29 (12.3) 5.46 (10.29) 7.82 (8.87) No-communication (25) Compens>Punish 20.0 15 (19.61) 14.40 (19.92) 0.6 (1.34) Punish> Compens 56.0 18.08 (5.84) 1.86 (3.13) 16.21 (5.21) Equal 24.0 8.33 (9.83) 4.17 (4.92) 4.17 (4.92) All 100 15.12 (10.92) 4.92 (10.04) 10.20 (8.33)
3.4. Altruistic punishment Vs. altruistic compensation between LE/HE and NC condition
3.5. Altruistic punishment Vs. altruistic compensation: Their relationship with the rest of our measures
Firstly, we see that the correlation coefficients between empathy and the two reactions, they have the hypothesized signs, namely positive for compensation (rs=.150) and negative
for punishment (rs=-.012), however, they are weak and statistically insignificant (see Table
4). Secondly, we observe a weak, negative, insignificant correlation between compensation and punishment points. Thirdly, there was a strong and significant correlation (rs=.577**)
between compensation points and the compensation question (compensate) whereas a weak and less significant one (rs=.241*) with the punishment question (punish). On the other
hand, punishment points were moderately and significantly (rs=.413**) correlated with the
punishment question whereas weakly and insignificantly with the compensation question. Those two pairs of correlations suggest indeed that increase points served mainly as compensation whereas the reduction points played the role of punishment as intended. Furthermore, based on both the means’ comparison analysis and the correlation analysis, we can detect four important differences between the choice of punishment and compensation; the intention of participants to maximize their points, the intention to help Player B, the feelings of guilt towards Player B and the participants’ hypothetical initial choice in case they were in Player A’s position.
The most significant difference between the choices of punishment versus compensation was in the participants’ intention to maximize their points. This conclusion can be derived from the comparison of the mean of the group that preferred compensation (M=2.86, SD=1.83) with the one that preferred punishment (M=4.98, SD=2.22) which appeared as the most significant among all, U=125.5, p= .002. Similarly, while maximize is the most correlated variable with compensation (rs=-.671**), the same correlation with punishment is
close to zero and with an opposite sign (rs=.085).
The second significant difference between the preference for compensation and punishment was in the participants’ intention to help Player B (M=4.93, SD=2.06 vs. M=3.58, SD=2.06),
U=175.5, p= .036. Again, although the correlation of compensation with help was strong and
significant (rs=.545**) the respective one for punishment is lower than .1 (rs=.095).
Similar correlation patterns hold for the feelings of guilt towards Player B and compensation (rs=.595**) versus punishment (rs=.002). The difference between the two relevant means
although is marginally insignificant with regards to the preference categorization (M=4.02, SD=1.16 vs. M=3.04, SD=1.77), U=185.5, p= .061, it is highly significant in the stricter group categorization for the ones that chose only compensation (M=3.87, SD=1.18) and only punishment (M=1.98, SD=1.061), U=26, p= .000. On the other hand, there is not a significant difference concerning the guilt towards Player A neither in the means of the two groups nor in the correlations.
15 be positively and almost strongly correlated with compensation (rs=.497**) whereas not
correlated at all with punishment (rs=.006).
Concerning the rest of the measures, there is neither any significant difference between the compensation and the punishment groups nor in their correlations.
4. Discussion
The aim of this study was to demonstrate the role of empathy on the choice between altruistic punishment and compensation as a reaction to an observed injustice. Empathy, although an established predictor of altruistic behavior, has received too little attention in the field of Economics. Moreover, we intended to indicate that altruistic compensation per se, although overlooked as a third party reaction to injustice, it comes together with altruistic punishment and fulfills different concerns.
Our results suggest that in cases of injustice, third party observers prefer to altruistically punish the violator rather than altruistically compensate the victim of the injustice. This is in accordance with previous studies on justice literature (Gromet & Darley, 2009). Yet, in a game setting similar to ours which also included the option to choose a combination of altruistic punishment and altruistic compensation, compensation was preferred over punishment (Lotz et al., 2011). Nevertheless, altruistic compensation proved to play an important role for restoring justice since it was not ignored as an available option from the
Table 4. Correlation matrix for the four main variables
Total points compensation punishment empathy anger responsibility guiltA guiltB Justice Total points 1.000 .331** .799** -.013 .150 .287** .332** .178 -.129 compensation .331** 1.000 -.126 .150 .329** .439** .345** .595** -.289** punishment .799** -.126 1.000 -.012 .136 .197 .297** .002 -.097 empathy -.013 .150 -.012 1.000 .337* .157 .182 .278* -.012
**. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-tailed).
Table 4 (cont.). Correlation matrix for the four main variables
offer violated help punish compensate equal maximize friendly selfish influence Total points .236* .296** .176 .372** .307** .322** -.231* -.107 .091 -.176 compensation .497** .366** .545** .241* .577** .486** -.671** -.217 .182 .070 punishment .006 .288** .095 .413** .146 .275* .085 .003 .082 -.203 empathy -.071 .183 .388** .205 .209 .199 .042 .098 .040 .095
majority of our participants. Moreover, we found that our participants as third party observers kept on average for themselves almost as many points as the victim ended up with after receiving the compensation points they assigned to him/her. In that way compensation contributed both in the equality among the three parties involved and in the one between the violator and the victim of the injustice.
From the comparison of participants’ reactions in the low-empathy condition with the ones in the high-empathy condition we conclude that there was a trend towards our hypothesis concerning the distribution of the assigned points. Although both low-empathy and high-empathy group exhibited similar levels of altruism and cared likewise about justice and equality, they chose different strategies for restoring them. The former showed a strong preference for punishment of the justice violator, whereas the latter used a substantial level of compensation for the victim which did not differ significantly from the level of punishment. Moreover, based on the same analysis, the no-communication condition resembled more to the low-empathy one. Therefore, the high-empathy condition remained the only one in which the amount and the frequency of compensation did not differ significantly with the respective ones of punishment.
17 be either to strictly follow Batson’s procedure or in case of similar constrains to ours to instruct the participants to also write a personal note in the beginning of the experiment as was tested by Rumble et al. (2010).
The second limitation of our study concerns the potential effect of internal competition in the points’ distribution. Budget restrictions lead us to introduce the monetary motivation of our participants in the form of lottery tickets as has been tested in previous studies (Nelissen & Zeelenberg, 2009) instead of endowing each participant with an initial amount of money as in the standard third party punishment game (Fehr & Fishbacher, 2004; Lotz et al., 2011; Leliveld et al., 2012). We suspect that this type of motivation brought up internal competition and affected participants’ reactions. Our suspicion is based both on the participants’ answers in the motivation open questions and the fact that a substantial percentage of punishment was concentrated in the levels between 16 and 20 points. With this level of punishment and provided there is zero compensation, participants end up with the most points among the three players. Some of those participants motivated indeed their punishment level by their desire to end up with more points than the justice violator (Player A) even if they had to decrease their overall chances for winning since the three lotteries were run among the whole sample of participants and not within each group. This confounded competitiveness effect was not expected since it did not appear in similar previous studies (Nelissen & Zeelenberg, 2009). A possible consequence of it might be the fact that there was a large preference for punishment and a small one for compensation compared to previous studies (Lotz et al., 2011; Leliveld et al., 2012) which probably decreased the statistical significance in the two groups’ differences. Another indication for the competitiveness effect is the fact that although there was a strong negative correlation between compensation and participants’ intention to maximize their points as someone would expect, the same correlation was slightly above zero for punishment. In that respect, it is important for future research to be cautious with this type of monetary motivation. If there are recourses restrictions as in our case, a potentially useful remedy might be to use one lottery for each participant with a smaller prize instead of one big lottery for the whole experiment.
tested in our study in cases of reactions to injustice. People who have a low-empathy trait might react very differently after being induced to feel empathic concern towards a victim of injustice compared to the ones with an already high-empathic trait. Another suggestion for future research would be to see these reactions to justice violation when the violator is not a physical person but an institution as occurs in many cases of real life experiences. It would be interesting even to explore how the reactions differ when empathic concern is not induced towards the victim but instead towards the violator of justice.
Overall, the results of our study concerning the different reactions to injustice by third party observers in the low-empathy condition compared to the high-empathy one, offer some very valuable messages. Firstly, we showed the importance of altruistic compensation in the reactions of unaffected third-party observers of injustice. Secondly, we indicated that moral emotions like empathic concern, which is as an established predictor of altruistic behavior, should not be ignored by economists in cases of third party reactions to injustice and in general. Lastly, the most important message is that inducing empathic concern towards the victim of an injustice can affect the type of reactions of third party observers willing to restore justice.
REFERENCES
Batson, C. D. (1991). The altruism question: Toward a social-psychological answer. Hillsdale, NJ: Lawrence Erlbaum Associates.
Batson, C. D., & Moran, T. (1999). Empathy-induced altruism in a Prisoner’s Dilemma.
European Journal of Social Psychology, 29, 909–924.
Batson, C. D., Kennedy, C. L., Nord, L. A., Stocks, E. L., Fleming, D. A., Marzette, C. M., Lishner, D. A., Hayes, R. E., Kolchinsky, L. M., & Zerger, T. (2007). Anger at unfairness: Is it moral outrage? European Journal of Social Psychology, 37, 1272–1285.
Bowles, S., & Herbert, G. (2006). Prosocial Emotions. In L. Blume, & S. Durlauf (Ed.), The
Economy as a Complex Evolving System III: Current Perspectives and Future Directions (pp.
339-366). Oxford and New York: Oxford University Press.
Bowles, S. (2008). Policies Designed for Self- Interested Citizens May Undermine ‘the Moral Sentiments’: Evidence from Economic Experiments. Science, 320, 1605–1609.
Brandts, J., & Charness, G. (2011). The strategy versus the direct-response method: A first survey of experimental comparisons. Experimental Economics, 14, 375–398.
Coke, J. S., Batson, C. D., & McDavis, K. (1978). Empathic mediation of helping: A two-stage model. Journal of Personality and Social Psychology, 36, 752–766.
19 Davis, M. H. (1994). Empathy: A social psychological approach. Madison, WI: Brown & Benchmark publishers.
de Waal, F. B. M. (2008). Putting the altruism back into altruism: The evolution of empathy.
Annual Review of Psychology, 59, 279−300.
Eisenberg, N. & Miller, P. A. (1987). The relation of empathy to prosocial and related behaviors. Psychological Bulletin, 101, 91-119.
Fehr, E. & Schmidt, K. M. (1999). A theory of fairness, competition, and cooperation.
Quarterly Journal of Economics, 114, 817–868.
Fehr, E., & Gächter, S. (2002). Altruistic punishment in humans. Nature, 415, 137–140. Fehr, E., Fischbacher, U., & Gächter, S. (2002). Strong reciprocity, human cooperation and the enforcement of social norms. Human Nature, 13, 1–25.
Fehr, E., & Fischbacher, U. (2003). The nature of human altruism. Nature, 425, 785–791. Fehr, E., & Fischbacher, U. (2004). Third-party punishment and social norms. Evolution and
Human Behavior, 25, 63–87.
Fontaine, P. (1997). Identification and economic behavior: sympathy and empathy in historical perspective. Economics and Philosophy, 13, 261–280.
Fontaine, P. (2001). The changing place of empathy in welfare economics. History of Political
Economics, 33, 387–409.
Frith, C. D., Singer, T. (2008). The role of social cognition in decision making. Philosophical
Transactions of the Royal Society B-Biological Sciences, 363, 3875–86.
Gintis, H., Bowles, S., Boyd, R., & Fehr, E. (2003). Explaining altruistic behavior in humans.
Evolution & Human Behavior, 24, 153–172.
Gromet, D. M., & Darley, J. M. (2009). Punishment and beyond: Achieving justice through the satisfaction of multiple goals. Law & Society Review, 43, 1–38.
Hechter, M., & Opp, K. D. (2001). Social norms. New York: Russell Sage Foundation.
Henrich, J., Boyd, R., Bowles, S., Camerer, C. F., Fehr, E., Gintis, H., McElreath, R., Alvard, M., Barr, A., Ensminger, J., Henrich, N. S., Hill, K., Gil-White, F., Gurven, M., Marlowe, F. W., Patton, J. Q. & Tracer, D. (2005). “Economic man” in cross-cultural perspective: Behavioral experiments in 15 small-scale societies. Behavioral and Brain Sciences, 28, 795–815; discussion 815–55.
Hoffman, M. L. (2000). Empathy and its Development: Implications for Caring and Justice. New York: Cambridge University Press.
Jones, E. E., & Pittman, T. S. (1982). Toward a general theory of strategic self-presentation. In
J. Suls (Ed.), Psychological perspectives on the self. Hillsdale, NJ: Erlbaum.
Kirman, A., Teschl, M., 2010. Selfish or selfless? The role of empathy in economics.
Philosophical Transactions of the Royal Society B-Biological Sciences, 365, 303–317.
Lamm, C, Batson, C. D. & Decety, J. (2007). The neural substrate of human empathy: effects of perspective-taking and cognitive appraisal. Journal of Cognitive Neuroscience, 19, 42-58. Leliveld, M.C., Dijk, E. van, & Beest, I. van (2012). Punishing and compensating others at your own expense: The role of empathic concern on reactions to distributive injustice. European
Journal of Social Psychology, 42, 135-140.
Loewenstein, G., & Small, D. A. (2007). The scarecrow and the tin man: the vicissitudes of human sympathy and caring. Review of General Psychology, 11, 112–126.
Lotz, S., Okimoto, T., Schlösser, T., & Fetchenhauer, D. (2011). Punitive versus compensatory reactions to injustice: Emotional antecedents to third-party interventions. Journal of
Experimental Social Psychology, 47, 477–480.
Nelissen, R. M. A., & Zeelenberg, M. (2009). Moral emotions as determinants of third-party punishment: Anger, guilt, and the functions of altruistic sanctions. Judgment and Decision
Making, 4, 543–553.
Orne, M. (1962). On the social psychology of the psychological experiment: With particular reference to demand characteristics and their implications. American Psychologist, 17, 776– 783.
Rumble, A. C., Van Lange, P. A. M., & Parks, C. D. (2010). The benefits of empathy: When empathy may sustain cooperation in social dilemmas. European Journal of Social Psychology,
40, 856−866.
Rabin, M.(1993). Incorporating fairness into game theory and economics. American
Economic Review, 83, 1281–-1302.
Smith, A. (1759/1976). The theory of moral sentiments. Oxford: Clarendon Press.
Singer, T., Seymour, B., O'Doherty, J., Kaube, H., Dolan, R. J. & Frith, C. D. (2004). Empathy for pain involves the affective but not sensory components of pain. Science, 303, 1157-1162. Singer, T.3 Kiebel, S. J., Winston, J. S., Dolan, R. J. & Frith, C. D. (2004). Brain responses to the acquired moral status of faces. Neuron, 41, 653-662.
Singer, T. & Fehr, E. (2005). The neuroeconomics of mind reading and empathy. American
21 Singer, T., Seymour, B., O'Doherty, J., Stephan, K. E., Dolan, R. J. & Frith, C. D. (2006). Empathic neural responses are modulated by the perceived fairness of others. Nature, 439, 466-469.
