Abstract
Why do victims sometimes not receive the help they need? One reason may be that people believe punishing perpetrators restores justice, which makes them less willing to help victims if the perpetrator has been punished. Participants who were first asked how much to punish a perpetrator subsequently recommended less compensation for the victim relative to participants who were asked about compensation first. In contrast, participants punished perpetrators to the same degree regardless of whether they were first asked about compensation (Study 1). These effects also held when a third party administered the initial response (Studies 2 and 3). Punishment increased people’s belief that justice had been restored, which decreased their desires for victim compensation (Study 3). Thus, the extent to which individuals are concerned about victims is influenced by whether they first consider perpetrator punishment.
In December 2008, notorious Wall Street financial advisor Bernie Madoff was convicted of defrauding investors of more than 50 billion dollars. People wanted him to be imprisoned to punish him for his crime and because they thought it would make his victims feel better about their loss (Fenton, Doyle, & Geller, 2009). However, sentencing Madoff to prison for a maximum of 150 years did nothing to financially compensate his victims for their losses. Indeed, as of December 2013, more than two thirds of the money has yet to be returned to his victims (Smith, 2013).
Why do victims sometimes not receive the help they deserve? Although there are often not enough resources to help victims, an additional possibility is that punishment restores people’s sense of justice to such a degree that it leads individuals to neglect victims’ needs. For example, Madoff’s victims may have been neglected because people felt that Madoff’s incarceration effectively restored justice. In this article, we test whether the increased sense of justice that individuals feel after learning perpetrators have been punished reduces their willingness to provide compensation to victims. In particular, we explore the asymmetrical relationship between third parties’ desires for punishment and compensation and posit that the presence of punishment decreases individuals’ willingness to compensate victims, whereas the presence of compensation does not affect individuals’ willingness to punish perpetrators.
Retributive and Utilitarian Goals
Transgressions threaten people’s belief that the world is just (Hafer & Bégue, 2005; Lerner, 1970; Lerner & Miller, 1978). As a result, victims and third parties are motivated to restore justice (Lerner, 2003) and engage in multiple strategies to do so, such as punishing, exacting revenge, demanding restitution, and appealing for compensation, among others.
People’s choice among these strategies is influenced by whether their goals are retributive or utilitarian. Whereas those with retributive goals are motivated to ensure that offenders receive their “just deserts” and suffer in proportion to the harm they caused (Carlsmith & Darley, 2008; Graham, Weiner, & Zucker, 1997), those with utilitarian goals (either in addition to or instead of retributive goals) punish to make the victim feel better or to make society safer (Carlsmith, Darley, & Robinson, 2002). Although they offer utilitarian justifications for punishing, people are primarily motivated by retribution (Carlsmith, 2006).
To meet these retributive and utilitarian goals, people may engage in multiple responses to wrongdoing, such as punishing perpetrators and compensating victims (Baron & Ritov, 1993; Darley & Pittman, 2003; Lotz, Okimoto, Schlösser, & Fetchenhauer, 2011; van Prooijen, 2010). Punishment involves intentionally sanctioning someone who has violated a law, rule, or norm (Vidmar & Miller, 1980). In contrast, compensation is a victim-focused strategy meant to restore victims to their pretransgression state (Darley & Pittman, 2003). It typically involves financial remuneration or repair for injury or damages incurred but may also take the form of “compensatory damages” above and beyond the cost of the wrongdoing. We argue that punishment and compensation are differentially able to satisfy people’s retributive and utilitarian concerns. Compensation may make victims feel better but does not do anything to punish perpetrators or benefit society. By comparison, punishing is meant to serve perpetrators their just deserts, make victims feel better, and ensure a safer society, thus meeting both retributive and utilitarian justice goals.
Punishment Reduces Compensation
Punishment effectively meets people’s retributive goals: In addition to serving the perpetrator his or her just deserts, it enables people to signal moral disapproval of the transgression and reassert their values that the perpetrator violated (Vidmar & Miller, 1980). People with retributive goals justify punishment as morally deserved by the offender (Newheiser, Sawaoka, & Dovidio, 2012). By selecting a magnitude of punishment that is proportional to the severity of the violation, people correct the imbalance in justice that the transgression created (Carlsmith et al., 2002; Darley & Pittman, 2003; Graham et al., 1997).
Punishment also serves utilitarian functions that benefit victims and society as a whole. First, sanctions deter perpetrators from recidivism and serve to incapacitate them for a period of time (e.g., prison sentences), thus physically preventing them from reoffending. Second, punishment may deter other would-be offenders from committing the same crime by signaling that such actions are socially unacceptable and will be similarly punished. People are most satisfied with punishment when offenders understand why sanctions are being imposed (Gollwitzer & Denzler, 2009), suggesting that people want punishment to send offenders a message that wrongdoing will not be tolerated. Indeed, people punish more when they know others will observe their actions (Kurzban, DeScioli, & O’Brien, 2007), implying that people understand punishment serves as a social signal. Finally, punishment can help victims feel better; it serves to reaffirm a victim’s status within a community (Miller, 2001; Vidmar & Miller, 1980; Wenzel & Thielmann, 2006) and can increase victims’ willingness to forgive perpetrators (Strelan & Van Prooijen, 2013).
In sum, punishment increases people’s sense that justice has been restored by fulfilling their retributive and utilitarian perpetrator-focused concerns and, to a lesser degree, their victim-focused concerns. As a consequence of greater perceived justice restoration, we predict that punishment will decrease third parties’ willingness to help victims.
Compensation and the Desire for Punishment
In contrast, we predict that third parties will continue to demand punishment even when victims have been compensated. Although compensation helps to restore victims to their pre-transgression state, thereby satisfying some utilitarian goals, it does not satisfy people’s desires for retribution (nor their utilitarian goal of making society a better place) which can only be accomplished through punishing. Furthermore, to the extent that people are concerned with deterrence, compensation only weakly signals intolerance of transgressions and does not incapacitate offenders; thus, it will do little to satisfy the “intuitive prosecutors” within people who desire to blame and punish offenders (Goldberg, Lerner, & Tetlock, 1999; Tetlock, 2002; Tetlock et al., 2007).
Compensation Versus Punishment
Other than one study indicating that children recommend less compensation for victims if the perpetrator has been punished (Miller & McCann, 1979), no research has systematically examined how punishment influences observers’ desires for victim compensation and vice versa. 1 Researchers have, however, investigated competing preferences for these two forms of justice. Although most theorists have argued that when confronted with injustice, people’s dominant concern is to punish and victim compensation is a secondary concern (Darley & Pittman, 2003; Lerner & Miller, 1978; Miller & McCann, 1979), the empirical evidence is more mixed. Some research indicates that people prefer punishing perpetrators to compensating victims (van Prooijen, 2010), whereas other research suggests that people prefer compensating to punishing (Lotz et al., 2011) and like individuals who compensate more than those who punish (Adams & Mullen, 2013). People’s preferences for punishment or compensation are undoubtedly influenced by a number of contextual factors such as the severity of the offense (e.g., Gromet & Darley, 2006) and whether the perpetrator intended harm (for a review, see Darley & Pittman, 2003; Mullen & Okimoto, in press). Our intention, however, was not to examine people’s competing preferences for punishment and compensation but rather to test how the provision of one element of justice restoration (e.g., punishment) influences desires for additional justice restoration strategies (e.g., compensation).
In three studies, we examine the influence of punishment on desires for compensation and the influence of compensation on punishment desires. In Study 1, we investigate whether third parties who are first asked about their desire to punish exhibit a lower desire for compensation than if they are first asked about compensation. In Studies 2 and 3, we investigate whether learning that a third party has assigned punishment (or compensation) decreases individuals’ desires for compensation (or punishment). In Study 3, we test whether the effect of punishment on reduced compensation is mediated by perceived justice restoration.
Study 1: Intentional Injury
We predicted that individuals who were first asked to make punishment recommendations would desire less compensation relative to participants who were asked about compensation first. In contrast, we predicted that punishment recommendations would not be influenced by whether people were first asked about compensation.
Method
Participants
Fifty-eight adults (53% women; M age = 35.26) participated in this study online for a chance to win a US$50.00 gift certificate to Amazon.com. They were recruited from a pool of people who had previously indicated their willingness to participate in research for scholars at a university.
Procedure
All participants read a scenario about Tony and Dylan, two coworkers at a tire factory who argued because Dylan was unwilling to trade shifts with Tony. To retaliate, Tony pulled a lever causing tires on an assembly line to arrive at Dylan’s workstation more quickly. The tires piled up and fell on Dylan, who sprained his shoulder and fractured his collarbone (see Online Supplement).
Participants were randomly assigned to answer items about how much they thought Tony should be punished and Dylan should be compensated in one of two orders. In the punish-first condition, participants answered the punishment items before the compensation items. In the compensate-first condition, participants answered the compensation items before the punishment items. The punishment and compensation items were presented on separate pages. Finally, participants responded to demographic measures.
Measures
All items were asked on 7-point scales ranging from not at all to to a great extent.
Punishment recommendations
We created a composite measure of punishment using 2 items: “To what extent should Tony be punished?” and “To what extent would punishing Tony be an appropriate response?” r(58) = .50, p < .001.
Compensation recommendations
We measured compensation using 2 items: “To what extent should Dylan be compensated?” and “To what extent would compensating Dylan be an appropriate response?” r(58) = .82, p < .001.
Results
We conducted a 2 (order: punish-first, compensate-first) × 2 (recommendation: punish, compensate) mixed analysis of variance (ANOVA) with recommendation as the repeated factor. Participants recommended punishing the perpetrator (M = 6.37, SD = 0.81) more than compensating the victim (M = 5.93, SD = 1.09), F(1, 56) = 7.78, p = .007. Moreover, the predicted interaction between order and recommendation was significant, F(1, 56) = 4.87, p = .03. Participants were less likely to recommend compensating when they were asked about compensation after punishment (M = 5.47, SD = 1.63) than before (M = 6.27, SD = 0.71), F(1, 56) = 5.71, p = .02, d = .67. Participants’ punishment recommendations did not differ as a function of whether they were asked about punishment before (M = 6.38, SD = 0.77) or after compensation (M = 6.36, SD = 0.88), F(1, 56) = .02, p = .90, d = .02.
Discussion
As predicted, people compensated the victim less if they made their compensation recommendation after rather than before punishing, but prior compensation recommendations did not influence punishment recommendations. One limitation of this study is that we did not specify the source of the compensation for participants; we address this shortcoming in Studies 2 and 3.
Studies 2A and 2B: Armed Robbery
In the United States, different levels of punishment can be meted out for the same crime because sentencing guidelines specify a range of penalties (e.g., a maximum of 15 years in prison) rather than a precise penalty. In addition, governments often play a role in restoring justice by prosecuting offenders or providing aid to victims who file a claim through the Office of Victim Services (OVS, 2012; http://www.ovs.ny.gov). Capitalizing on these procedures, in Study 2A, we varied the amount of punishment that a perpetrator received from government authorities to test for a negative linear relationship between the amount of punishment given and the amount of compensation participants would recommend victims receive. In Study 2B, we varied the amount of compensation that a victim received from government authorities to test whether it influenced participants’ punishment recommendations; we predicted it would not. 2
Method
Participants
Participants were 146 (64% female, M age = 35.77 years) and 120 (70.8% female, M age = 35.02 years) adults who received a US$3 Amazon.com gift card for their participation in Studies 2A and 2B, respectively. Participants could only complete one of the studies. They were recruited in the same manner as in Study 1.
Procedure
All participants read a scenario about Victor, the victim of an armed robbery. Peter, the offender, held a knife to Victor’s throat and took his wallet that contained US$100. Victor was stabbed but survived the attack. Peter was eventually apprehended but had already spent the stolen money (see Online Supplement).
In Study 2A, after reading about the transgression, participants were randomly assigned to one of four punishment conditions. They read either that (a) Peter’s case had not yet been brought before a judge or they were told that the typical punishment for crimes of this type ranges from 5 to 25 years in prison and that (b) Peter was convicted and put on probation; (c) Peter was convicted and sentenced to spend 5 years in prison; or (d) Peter was convicted and sentenced to spend 25 years in prison. After reading about Peter’s punishment, participants received information that the OVS helps victims and that compensation amounts typically range from the value of what was stolen to larger amounts that include money for associated attorney fees and emotional distress. Participants then responded to items measuring their compensation desires and demographics.
In Study 2B, after reading about the transgression, participants were randomly assigned to one of four compensation conditions. They read either that (a) Victor’s request for compensation had not yet been processed by the OVS or they learned that the typical amount of compensation for crimes like these ranges from the value of what was stolen to additional money for attorney fees and emotional distress or damages and that (b) Victor received US$25 but no compensation for attorney fees, emotional distress, or damages; (c) Victor received US$100 but no compensation for attorney fees, emotional distress, or damages; or (d) Victor received US$100 and the OVS paid his attorney fees and awarded him money for emotional distress and damages. Finally, participants learned that the district attorney could charge Peter with armed robbery, a Class B violent felony and that the typical punishment ranges from 5 to 25 years. Participants then responded to items measuring their punishment desires and demographics.
Measures
All items below were asked on 7-point scales ranging from strongly disagree to strongly agree.
Study 2A: Compensation
We created a composite of 2 items measuring participants’ desires to have the OVS compensate Victor: “I think Victor should be compensated by the OVS” and “Compensating Victor would be an appropriate response from the OVS,” r(146) = .86, p < .001.
Study 2B: Punishment
We used 2 items about participants’ desires to punish: “I think Peter should be punished” and “Punishing Peter would be an appropriate response,” r(120) = .76, p < .001.
Results and Discussion
Study 2A
An ANOVA revealed a significant effect of punishment magnitude on participants’ desires for compensation, F(3, 142) = 3.10, p = .029, ηp 2 = .06. A linear trend analysis demonstrated that the more Peter was punished, the less participants wanted to compensate Victor, t(142) = 2.55, p = .012 (see Table 1).
Mean Levels of Participants’ Compensation Desires as a Function of Perpetrator Punishment, Study 2A.
Note. Means with different subscripts differ significantly at p < .05. Standard deviations are provided in parentheses.
Study 2B
An ANOVA revealed there was not a significant effect of compensation magnitude on participants’ desires for punishment, F(3, 116) = 1.49, p = .221, ηp 2 = .04 (see Table 2).
Mean Levels of Participants’ Punishment Desires as a Function of Victim Compensation, Study 2B.
Note. Means with identical subscripts do not differ from one another. Standard deviations are provided in parentheses.
In summary, the more punishment a perpetrator received, the less likely participants were to recommend that the victim be compensated. Varying victim compensation amounts did not significantly affect participants’ punishment recommendations.
Studies 3A and 3B: Intentional Assault
Individuals might award less compensation to a victim after punishing an offender if they believe justice has been restored after punishment. Thus, in Study 3, we test whether perceptions of justice restoration mediate the relationship between punishment and decreased compensation. Moreover, we more closely simulate the decisions jurors and judges make by asking an additional question about the actual amounts of money that participants are willing to fine perpetrators (punish) and award to victims (compensate). Finally, we purposefully manipulated token amounts of punishment (and compensation) to ascertain whether even small, symbolic amounts of punishment would increase perceptions of justice restoration and decrease desires for compensation, thus providing a more conservative test of our hypotheses.
Method
Participants
Ninety-six participants (33% women; M age = 31.2) completed Study 3A and 95 participants (39% women; M age = 31.7) completed Study 3B via Amazon.com’s Mechanical Turk for US$0.40.
Procedure and Measures
All participants read a story about Victor, an employee who was promoted to project manager over Peter, another employee. Peter was upset that he had not been promoted, so he assaulted Victor as Victor left work; Victor suffered a concussion, a broken nose, and a bruised ribcage (see Online Supplement).
In Study 3A, participants were randomly assigned to one of two conditions. In the baseline condition, participants learned Peter was not punished due to the violation of a technical rule about courtroom proceedings. In the punishment condition, participants read Peter was convicted of assault and battery and had to perform 10 hr of community service as punishment.
In Study 3B, participants were randomly assigned to one of two conditions. In the baseline condition, participants learned Victor was not compensated due to budget cuts. In the compensation condition, participants read a description of the state’s Office for Victims of Crime (OVC, 2012: http://www.ojp.usdoj.gov/ovc/), and learned the OVC gave Victor US$100.00 but nothing for medical bills or attorney fees.
Finally, in both studies, participants were asked how much they thought justice had been restored, how much they wanted to compensate (Study 3A) or punish (Study 3B), and responded to demographic measures.
Studies 3A and 3B: Justice restoration
To measure how much participants thought justice had been restored following the state’s response, we asked participants 2 items, “To what extent was justice served?” and “To what extent did the state’s response restore your sense of justice?” on 7-point scales ranging from not at all to to a great extent. We combined these items into composite measures of justice restoration, r(96) = .82, p < .001 for Study 3A and r(95) = .87, p < .001 for Study 3B.
Study 3A: Compensation
Participants were told Victor could be awarded up to US$10,000 from the state’s OVC and were asked, “How much money should Victor receive?” on an 11-point scale ranging from US$0 to US$10,000 in increments of US$1,000. They also responded to the 2 compensation items from Study 1. Responses to all 3 items were standardized and combined into a composite measure of compensation, α = .86.
Study 3B: Punishment
Participants were told Peter could be fined up to US$10,000 if convicted of a felony charge of assault and battery, which would go to the state, and were asked, “How much money should Peter be fined?” on the same scale used previously. They also responded to the 2 punishment items from Study 1. Responses were standardized and combined into a composite measure of punishment, α = .83.
Results: Study 3A
Compensation
Participants compensated the victim less when the perpetrator had been punished (M = −0.20, SD = 0.82) than when the perpetrator had not been punished (M = 0.20, SD = 0.91), t(94) = 2.22, p = .029, d = .46.
Justice Restoration
Participants thought justice was restored more when the perpetrator was punished (M = 2.41, SD = 1.52) than when he was not punished (M = 1.51, SD = .86), t(72.35) = 3.57, p = .001, d = .76. 3 Moreover, increased justice restoration beliefs were associated with less compensation, B = −.17, SE(B) = .07, β = −.25, t(94) = −2.47, p = .015. We tested mediation with a bootstrapping strategy with 1,000 samples (Preacher & Hayes, 2004). The bootstrap coefficient for the indirect effect was −.14, and the 95% bias corrected confidence interval excluded zero [−.25, −.03]; thus, justice restoration mediated the effect of punishment on decreased compensation (see Figure 1).

The effect of third-party punishment on decreased compensation awards is mediated by perceived justice restoration. Note. Punishment was coded 0 (no response) and 1 (community service). Standardized βs are reported (the β in parentheses indicates the effect of punishment on compensation prior to controlling for justice restoration). **p ≤ .01. *p < .05.
Results: Study 3B
Punishment
There was not a significant effect of compensation on punishment, t(93) = .88, p = .38, d = .19. Participants punished similarly in the compensation (M = −0.08, SD = 0.93) and baseline conditions (M = 0.08, SD = 0.80).
Justice Restoration
Unexpectedly, participants perceived justice to have been restored less in the compensation (M = 1.71, SD = .86) than in the baseline condition (M = 2.52, SD = 1.71), t(76.10) = 3.20, p = .002, d = .67. However, perceived justice restoration was not significantly correlated with punishment, r(95) = .05, p = .625.
Discussion
Victims receive less help even when there is only a token amount of third-party punishment, because punishing increases people’s sense that justice has been restored. Participants more strongly believed that justice had been restored and therefore awarded less money to the victim when the perpetrator had been merely assigned to perform community service for assaulting a coworker relative to when the perpetrator had not been punished. However, participants punished the perpetrator the same amount regardless of whether the victim had received token compensation; it seems that compensation in the absence of punishment is not enough to provide observers with the sense that justice has been adequately restored.
General Discussion
We document asymmetrical effects of punishment and compensation on desires for additional responses to transgressions: People’s desires to help victims seem to be a function of whether they have first endorsed (or observed) punishment for perpetrators. Punishment reduced participants’ willingness to compensate victims relative to conditions where the perpetrator had not been punished. In contrast, victim compensation did not influence desires to punish perpetrators. Even minimal punishment increased participants’ perceptions that justice had been restored relative to not punishing, and this increased belief that justice had been restored led to decreased desires to compensate.
Punishment and compensation are frequent and often simultaneous responses to crimes. Whereas past research has explicated people’s preferences for one versus the other when presented simultaneously, the current research directly compares the effects of one on people’s willingness to provide the other. We find that punishment and compensation are not mutually substitutable for one another. More importantly, our research elucidates one reason why victims sometimes do not receive the attention they deserve after a crime has been committed. The legal systems in many countries ensure that crimes are brought first to criminal courts before victims can receive compensation in civil courts. Thus, careful consideration should be given to avoid any biasing effects of knowledge of prior punishment on compensation judgments.
Limitations and Future Directions
Although we found that compensation did not influence punishment, certain characteristics of the transgression might moderate this effect. For example, Darley and Pittman (2003) noted that desires for retributive justice increase with the severity of the transgression as well as the extent to which the perpetrator intended harm. For injustices less severe than the armed robbery and assault detailed in our studies (e.g., cases involving negligence or accidental harm), compensating victims might decrease the desire to punish (but see Gromet, Okimoto, Wenzel, & Darley, 2012). Future research should explore this possibility.
Consistent with the notion that transgression severity might influence the relationship between compensation and punishment desires, van de Calseyde, Keren, and Zeelenberg (2013) found that people punished perpetrators less when they transgressed against an insured victim relative to an uninsured victim, because they perceived the crime to be less severe and less unethical. For example, participants punished thieves less when they stole a camera from a victim who had purchased theft insurance for the camera relative to an uninsured victim. The difference between our results and van de Calseyde et al.’s raises the interesting possibility that third-party compensation (which may highlight the harm to the victim) may not reduce the perceived severity of the transgression in the same manner as an insurance payout. Further research should investigate whether the magnitude of third-party compensation (or punishment) influences the perceived severity of the transgression and people’s willingness to engage in additional justice restoration measures. 4
Finally, research could explore factors that moderate the effect of punishment on decreased compensation desires. For example, reminding observers that victims need help even after perpetrators are punished (e.g., by highlighting that victims are still suffering) might motivate observers to take additional action. Additionally, studies that more vividly elicit a sense of injustice (beyond that highlighted by the vignettes used in the present research) might more strongly evoke sympathy for victims (Jenni & Loewenstein, 1997; Lerner, 2003), thus more strongly motivating people to help victims even after perpetrators have been punished. Such studies would also extend the external validity of the present research. Finally, although our research suggests that observers believe punishing addresses victims’ concerns, future research should also address whether victims feel punishment adequately restores justice.
Conclusion
Compensation seems to require punishment to fully meet the goals of restoring justice in the eyes of third parties, whereas punishment seems to be adequate independent of compensation. Although punishment might facilitate forgiveness (Strelan & Van Prooijen, 2013; Tripp, Bies, & Aquino, 2007), the potential for further harm to victims is clear; having already been transgressed against, victims are less likely to receive compensation if third parties make it their priority to punish. Thus, attempts should be made to ensure that justice is restored in the eyes of victims and observers alike.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors received financial support for this research from their respective academic institutions.
Notes
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
