Victim impact testimony and the psychology of punishment.
INTRODUCTION I. THE SOCIAL PSYCHOLOGY OF PUNISHMENT: THE EVIDENCE THAT PEOPLE ARE GENERALLY MOTIVATED BY HARM-BASED RETRIBUTION II. VICTIM IMPACT STATEMENTS AS SUBJECTIVE ADVENTITIOUS HARM A. Background: "The Jurisprudence of Victimhood" B. The Unique Nature of Victim Impact Statements III. EXPERIMENTAL EVIDENCE FOR THE INFLUENCE OF SUBJECTIVE EMOTIONAL HARM ON PUNISHMENT JUDGMENTS A. Background: Previous Empirical Research B. The Current Study 1. Participants and Procedure 2. Materials and Design 3. Results IV. SOCIAL PSYCHOLOGICAL PROCESSES THAT ACCOUNT FOR THE INFLUENCE OF VICTIM IMPACT EVIDENCE ON PUNISHMENT JUDGMENTS A. Outcome Bias: Emotional Harm as a Heuristic for Culpability B. Decision-Maker Emotion C. Victim Impact Statements as Clinical Interviews: More Information Is Not Always Better CONCLUSION: REDUCING ARBITRARINESS IN CAPITAL SENTENCING JUDGMENTS APPENDIX The happy or unprosperous event of any action, is not only apt to give us a good or bad opinion of the prudence with which it was conducted, but almost always too animates our gratitude or resentment, our sense of the merit or demerit of the design. --Adam Smith (1)
When people make decisions about blame and punishment, harm matters. Adam Smith captured a powerful intuition when he asserted that, as a factual matter about human judgment, an unfortunate outcome prompts resentment, whereas a fortunate outcome prompts gratitude, and our judgment of an actor's culpability rests on the fortuity of the outcome. (2) Smith invites us to consider a person who hurls a large stone over a wall into a public street, without regard to where it might fall. (3) If the stone should accidentally kill a man, severe punishment (even the death penalty) for the stone hurler would be consistent with our natural sentiments. But it would be "shocking to our natural sense of equity ... to bring a man to the scaffold merely for having thrown a stone carelessly into the street without hurting any body." (4)
The criminal law, too, recognizes that harm matters, even when the harm in question is unforeseeable or adventitious. In most jurisdictions, for example, attempted murder and murder are punished differently. (5) Yet, whether a shooting victim dies or lives might depend on how quickly the ambulance arrived, whether the bullet hit or missed a vital organ, the skill of the surgeons who happened to be on call at the hospital, the victim's general health, and a host of other circumstances over which the perpetrator has no control and which he could not have foreseen. In this sense, the criminal law seems to track closely our intuition that blame and punishment should correspond to harm, even if the harm is unforeseeable or adventitious.
But is it really the case that intuitions about punishment and blame track harm at all times and under all circumstances? The distinction between attempted and completed crimes is just one example of the many ways in which the adventitious consequences of harm might (or might not) affect our sentiments about culpability and punishment. For example, does the type of harm (e.g., physical, financial, emotional) matter? Is the identity of the person who suffers the harm (e.g., a nurse or a homeless drug addict) important? Is the remoteness of the crime's harmful consequences (e.g., emotional suffering or stress that eventually leads to a heart attack) significant?
Ultimately, these are empirical questions. (6) The answers to these questions directly influence criminal sentencing generally, and capital punishment in particular, resulting in grave and irreversible consequences. The role that harm may or may not play in punishment judgments has been the subject of recent debate regarding victim impact testimony. Following the Supreme Court's decision in Payne v. Tennessee (7) to permit surviving friends and family of homicide victims to testify about the character of the victim and the emotional and psychological impact of the victim's death, commentators have criticized this type of testimony for its effects on capital sentencing. (8) In particular, there is a concern that the use of victim impact statements will introduce arbitrariness into the capital sentencing process and will represent a step backwards from Furman v. Georgia (9) and Gregg v. Georgia, (10) both of which emphasized the need for consistency in capital sentencing systems. (11)
The harms described in victim impact statements raise a variety of concerns about the adventitious nature of these statements and their impact on the decision about whether the defendant will ultimately live or die. First, commentators are particularly concerned that victim impact statements highlight the perceived relative worth of the victim, and consequently that the jury's judgment about whether to impose the death sentence will be influenced by this inappropriate factor. (12) In this sense, whether someone kills a person of high or low standing is adventitious because the death reflects neither the murderer's mental state nor the morality of the act itself. (13) In addition, victim impact statements detail the various harms that befall the victim or the victim's family after a crime, which might include financial complications resulting from a crime, stress-induced illnesses, and various other consequences that a defendant might not have been able to foresee at the time of the crime. (14) In this Article, we focus on one specific consequence that is very often the subject of victim impact statements: emotional harm to victims or survivors. We isolate this issue because, compared to other types of harms, it is particularly subjective, amorphous, and likely to vary across individuals and circumstances. In short, the emotional fallout of a crime to a victim or a victim's family is a paradigmatic example of adventitious harm.
Does evidence about emotional harm inform people's judgments about punishment? In this Article, we report our own empirical findings from a new experiment suggesting that adventitious, emotional harm affects jurors' sentencing decisions. In discussing our findings, we consider a number of psychological principles that may account for people's reliance, at least in part, on emotional harm as a factor in determining punishment. Because these are general principles, we have no reason to believe that they are particular to one type of crime or another. Therefore, although our study and much of our discussion consider the general influence of emotional harm information on punishment, we end by considering its consequences in capital sentencing in particular.
In Part I of this Article, we discuss the role of retribution in the psychology of punishment. In Part II, we examine the nature of accidental harm in the psychology of punishment and victim impact statements as a species of such accidental harm. In Part III, we present empirical evidence that suggests that victim impact statements do in fact influence lay sentencing decisions. After discussing the results of this empirical research, in Part IV we examine several social psychological explanations for the influence of adventitious harm in general, and victim impact statements in particular, on punishment judgments. Finally, we conclude by discussing recommendations for reducing the arbitrariness produced by the introduction of victim impact statements in capital sentencing hearings.
THE SOCIAL PSYCHOLOGY OF PUNISHMENT: THE EVIDENCE THAT PEOPLE ARE GENERALLY MOTIVATED BY HARM-BASED RETRIBUTION
Although philosophers have identified a number of different goals associated with punishing criminals--including deterring the offender from committing future crime, deterring others from crime, isolating and incapacitating the offender, rehabilitating the offender, and retribution (15)--there is a separate empirical question about the extent to which these or other factors actually motivate laypersons' punishment preferences. Empirical research on the psychology of justice supports an emerging consensus that people's punishment judgments are guided to a large degree by a harm-based retributive psychology. Some research, for example, has examined the extent to which laypersons' ideas about punishment comport with federal sentencing guidelines, in order to learn whether the guidelines are at odds with lay intuitions about justice. (16) Similarly, Paul Robinson and John Darley have studied the extent to which laypersons' punishment judgments accurately reflect criminal law principles embodied in state statutes and the Model Penal Code. (17) Other research has focused on the consistency between laypersons' punishment intuitions and legal principles regarding punitive damages in civil cases. (18) In most instances, people's punishment judgments track the harm caused by the crime. For example, laypeople and judges alike typically believe that a person who steals five dollars from a church poor box should be punished less severely than a person who embezzles fifty thousand dollars from the same church. (19) The fact that judges pay particular attention to harm in determining punishment judgments is evidenced in their own descriptions of their sentencing practices for white-collar crimes. (20)
Thus, as an empirical matter, people generally are motivated to punish so that the offender suffers as the offender has made others suffer, (21) and to impose the punishment that the offender deserves for the wrong that he or she has committed. (22) In other words, punishment decisions generally reflect a moral reaction to a wrong--a sense of deservingness. (23) This motivation to punish based on the moral nature of the offense appears to be independent of the punishment's potential to satisfy other sentencing goals such as deterrence. (24) Indeed, as a psychological matter, a retributive philosophy has been described as the "default" sentencing strategy. (25) That is, when people are instructed to punish based on retribution, their sentences correspond most closely with sentences given without explicit instructions on strategy. (26) Finally, when asked to rank the philosophies they believe should guide sentencing, people explicitly choose retribution as the most important principle. (27)
As demonstrated in research on people's reactions to accidents, attention to harm in determining punishment judgments sometimes includes attention to adventitious harm. For example, the more severe an accident victim's injury, the more blame and responsibility people attribute to the person who caused the accident. (28) One explanation for this phenomenon is that people are motivated to make defensive attributions. (29) In other words, when harm is severe, people are threatened by the possibility that a similarly serious injury could befall them. In order to reassure themselves that such severe outcomes do not occur accidentally or randomly, people assign greater responsibility for the harm to the relevant actors. (30) For example, a driver whose unoccupied parked car rolls down a hill is judged to be more responsible and more blameworthy when the car happens to hit and injure a group of pedestrians than when the same car happens to hit a tree and injures no one. People assign responsibility and blame in this manner despite the fact that in both scenarios, the driver's state of mind and behavior were identical.
The severity of harm therefore influences judgments of responsibility and blame. Moreover, severity affects judgments about adventitiousness itself. In particular, people view adventitious consequences as less adventitious when the harm is serious than when the harm is minor. This tendency is illustrated in a hypothetical in which a bank robber's bullet misses its intended target (the teller who is pressing the alarm), and hits a bank customer instead. (31) Under these circumstances, people are less willing to acknowledge the chance nature of the harm to the customer when the harm is severe (victim is paralyzed) than when the harm is mild (victim is grazed by bullet). (32) We find it threatening to believe that a severe outcome could be due to chance because that would imply that it could happen to us. We therefore reduce the threat of accidental harms by viewing the event as non-accidental and assigning more responsibility to the agent of the harm. (33) Evidence from a recent meta-analysis of existing studies forcefully demonstrates that the more severe the harm resulting from accidents, the greater the perceived responsibility, the greater the blame for the accident, and the larger the mock jurors' monetary damage awards. (34)
This review suggests two conclusions: harm is a critical factor in people's views about just punishments, and harm is not rendered irrelevant simply because it is adventitious or unforeseen. Given these conclusions, it seems likely that a crime victim's ability to cope emotionally with a crime will influence jurors' ideas about appropriate punishment. Before addressing this question empirically, we discuss the unique nature of emotion-based victim impact evidence and the current status of victim impact statements in criminal trials.
VICTIM IMPACT STATEMENTS AS SUBJECTIVE ADVENTITIOUS HARM
A. Background: "The Jurisprudence of Victimhood" (35)
In Booth v. Maryland, the U.S. Supreme Court considered whether, in a capital sentencing hearing, the prosecution's presentation of information about the murder victim's family members' opinions about the crime and about the defendant violated the Eighth Amendment. (36) The Court concluded that presentation of victim impact statements in a capital sentencing hearing creates an unacceptable risk that the jury may impose the death penalty in an arbitrary and capricious manner, because such evidence diverts the jury's attention away from the crime and the defendant and toward the character of the victim and the crime's effect on his family. (37) Therefore, the Court noted, "[a]llowing the jury to rely on a [victim impact statement] ... could result in imposing the death sentence because of factors about which the defendant was unaware, and that were irrelevant to the decision to kill." (38) Finding this possibility to be unacceptable, the Court held that the introduction of victim impact statements in a capital trial violates the Eighth Amendment. (39)
Only four years later, however, the Court reconsidered its position on the constitutionality of victim impact statements. In Payne v. Tennessee, the State of Tennessee introduced victim impact statements in the penalty phase of a capital trial in clear violation of the Court's holding in Booth. (40) In a stark reversal of its decision in Booth, the Court overruled itself and held that juries may consider victim impact statements in deciding whether to sentence a defendant to death. (41) Specifically, the Court permitted the victim's family and friends to testify about the victim's personal characteristics (42) (e.g., "she enjoyed playing bridge and had many friends") and the crime's psychological impact on the survivors (e.g., "I still cry every day, even though this crime happened two years ago"). (43)
The Court's decision in Payne was a major victory for a then-decade-long movement to empower victims in the criminal justice system. Under President Ronald Reagan, the President's Task Force on Victims of Crime issued a report concluding that a serious imbalance existed between the rights of defendants and the rights of victims, and proposed a constitutional amendment to allow victims "to be present and to be heard at all critical stages of judicial proceedings." (44) The political momentum and influence of the victims' rights movement continues to intensify even after Payne, and the wisdom of a constitutional amendment for victims' rights has been debated in recent congressional sessions. (45) Statutes in all fifty states, as well as the U.S. Code, provide for some form of victims' rights in criminal proceedings generally, (46) and at least thirty-two of the thirty-eight death penalty states permit victim impact statements in the penalty phase of capital trials. (47)
The Payne decision was at once both wholly consistent with the tenor and goals of the victims' rights movement and highly controversial. (48) The Court itself was bitterly divided, in part because Payne's holding overturned two prior decisions on the issue. (49) In addition, the majority and dissenting opinions in Payne reveal two radically different views of the appropriate goals of sentencing and of the type of information that juries should consider when deciding whether to impose the death penalty. Writing for the Payne majority, Chief Justice Rehnquist identified several functions that victim impact statements serve in the penalty phase of a capital trial. First, the majority contended that victim impact statements reflect the true level of harm caused by the criminal offense. (50) Criminal law takes harm into account when apportioning responsibility, even when the harm is partly attributable to adventitious factors. For example, a successful assassin bears more legal responsibility than an unsuccessful one whose bullet is deflected. (51) Therefore, a capital jury should likewise consider harm caused by the offense when deciding whether to impose death as the penalty. The Payne majority also argued that victim impact statements serve a balancing function. Because courts permit defendants to present mitigating evidence about their background and character in the penalty phase, victim impact statements are necessary to provide the jury with an opportunity to consider the victim as a unique individual, rather than as a "`faceless stranger.'" (52)
Justice Stevens's dissent, in contrast, raised several concerns about the propriety of permitting victim impact statements in capital sentencing hearings, three of which are most relevant here. First, Justice Stevens contended that the sole purpose of victim impact statements is to incite the emotion of the jury, and that as a consequence jurors will base their decision whether to impose the death penalty on emotion rather than reason. (53) Second, achieving a balance between prosecution and defense is not among the goals of criminal prosecution. Because of this, and because the victim is not a party in criminal proceedings, the dissent reasoned that no justification exists for the majority's position that victim impact statements serve a legitimate balancing purpose. (54) Third, victim impact statements introduce evidence of harm that is unforeseeable and not clearly identified until after the crime, and therefore is irrelevant to the jury's punishment judgment. (55)
The 6-3 decision in Payne, and the strong language contained in the various concurring and dissenting opinions, reflect a sharp division in the Court regarding the philosophical and legal justifications for basing punishment judgments on harms that are subjective and, arguably, unforeseen. Should judgments of a person's praiseworthiness or blameworthiness for a given action depend on factors or conditions that are beyond that person's control? In the next subpart we argue that victim impact statements present a unique harm that warrants an analysis unlike that applied by the courts to more traditional harms.
B. The Unique Nature of Victim Impact Statements
Victim impact statements describe harm that is at once highly emotional, adventitious, and causally remote. The harm described by victim impact statements is adventitious in the sense that in many cases the victim's character and the nature of the reaction of the victim's family are facts that the defendant has no knowledge of at the time of the crime. In addition, the harm described by victim impact statements, while undoubtedly very real, is far down the causal chain from the defendant's criminal act. For example, in capital sentencing hearings after Payne, trial courts have admitted testimony from family members concerning medical developments, such as heart attacks, that they have experienced since the crime. (56) Other courts have admitted victim impact statements that describe how the crime has affected relationships within the murder victim's extended family, such as one statement that explained how the decedent's sister's marriage suffered and ended in divorce. (57) While most of the ill effects that are now routinely admitted in capital cases are undoubtedly causally related to the crime, the causal relationship is, in some instances, quite remote. This fact raises the question of how far down the causal chain courts should be willing to go when deciding what kind of victim testimony to admit in the penalty phase of a capital trial. (58)
The Payne Court held that, as a category of evidence, victim impact statements are relevant to the jury's decision whether to impose the death penalty. (59) Yet, relevance as an evidentiary matter does not guarantee that the evidence will persuade the jury, and empirical evidence on this question is sparse. Empirical investigations regarding the link between magnitude of harm and severity of punishment typically have examined physical or otherwise objective harms sustained by a victim (e.g., the type of physical injury sustained or the amount of money stolen). (60) On the other hand, what makes victim impact statements particularly intriguing for understanding the psychology of retribution is precisely what contributes to its controversial nature: compared to other types of harms, emotional injuries are less quantifiable and necessitate subjective appraisals about their significance. (61) If harm is expressed in the form of a victim's subjective, emotional reactions to a crime, tension will exist between placing blame based on the level of harm and recognizing that harm which is unforeseen, adventitious, or causally remote might merit less blame. This tension is reflected in laypersons' intuitions about the factors that should be considered in the determination of just punishments. Although the psychology of punishment--the understanding of people's motives and goals for responses to wrongdoing--has burgeoned within social psychology in recent years, (62) jury responses to victims are an understudied portion of the empirical literature. In the next Part, we discuss what we do know empirically about the influence of victim impact statements on punishment judgments.
EXPERIMENTAL EVIDENCE FOR THE INFLUENCE OF SUBJECTIVE EMOTIONAL HARM ON PUNISHMENT JUDGMENTS
A. Background: Previous Empirical Research
What do we know about the effects of subjective harms, such as those expressed in victim impact statements, on culpability and punishment judgments? The existing body of work in this area is quite limited, but appears to be consistent with the general principle that victim impact statements which communicate greater harm produce more severe punishment judgments, at least in controlled, experimental simulations. A handful of studies have examined victim impact evidence in capital case simulations, although two of these studies specifically avoided examining sentencing per se, reflecting the well-grounded fear that a simulation would not adequately represent the intensity of emotion in an actual death penalty sentencing hearing. (63) Nonetheless, each of these studies is instructive. In both studies, the researchers varied the victims' character and found that mock jurors perceived the crime as more serious, (64) and the survivors as having suffered more psychological, physical, and financial harm, when the victim was portrayed in a more positive light. (65)
In a different study, undergraduate students acted as mock jurors and made capital punishment decisions. (66) The results of that study showed that mock jurors were more likely to vote for a death sentence when victim impact evidence was present rather than absent, and this was especially true when the crime involved aggravating circumstances. (67) Another simulation using student mock jurors also found increased recommendations for death sentences in the presence of a victim impact statement, but this was true only for post-deliberation individual verdicts, not for mock jury group verdicts. (68)
Experimental studies on the use and effect of victim impact statements in noncapital cases are rare. In one study, researchers asked Australian adults to read case files about a robbery or a rape, which varied in their description of how well the victim was coping subsequent to the crime. (69) Even though victim-coping is an adventitious feature of the crime, (70) subjects who read about a victim who was coping poorly were more punitive in sentencing than were subjects who read about a victim who was coping well. (71) These results, however, are difficult to interpret, because the method used in the study confounded the information source with the type of harm. When the victim coped poorly, the prosecutor provided the victim information; when the victim coped well, the defense provided the victim information. (72)
In another study, researchers presented mock jurors with a transcript of victim testimony interspersed with written descriptions of the victim's emotional display. (73) The results indicated that a strong emotional display by the victim led observers to view the victim in a more positive light, which in turn resulted in harsher punishment for the defendant. (74) Interestingly, the study also found that emotional displays by the victim and by the defendant predicted subjects' impressions of the original offense, which in turn also predicted punishment judgments. (75) This study, however, was limited to a great extent by its use of undergraduates as mock jurors and by a rather impoverished method of conveying information about the victim's emotional display. (76)
Although the preceding experimental studies suggest that adventitious harms can affect laypersons' views of appropriate sentences in both capital and noncapital simulation contexts, (77) each study contains factors that make it difficult to assert this proposition without reservation. The studies involving capital cases either had mixed results (78) or opted not to examine sentencing at all. (79) The studies involving noncapital crimes described emotional harm in confounded (80) or problematic ways. (81)
In order to further examine the question of emotional harm as a factor in punishment judgments, we designed a simple experiment. In this experiment, we randomly assigned participants to three groups. The only difference between the three groups was the description of the severity of the victim's emotional response to the crime. Because capital sentencing is a complicated punishment decision that is very difficult to simulate both procedurally and emotionally, we opted to study the effects of victim impact statements on relatively common types of crimes.
In the subpart that follows, we describe the methodology and present the results from our initial study of the effect of victims' emotional suffering on laypersons' judgments about criminal punishment. Specifically, this study tests the hypothesis that the severity of emotional harm described by a victim in a criminal proceeding is directly related to the severity of the sentence imposed on the defendant.
B. The Current Study
1. Participants and Procedure
Three hundred and two adults volunteered to participate in the study. (82) Participants were recruited to fill out a brief questionnaire during the lunch hour in the lobby of a busy university administration building in downtown Chicago. (83) Participants varied widely along various demographic features. Only 36% of the participants were students; (84) the remaining participants worked in nearby offices or were members of the public who happened to be passing through the lobby. Thirty-three percent of the participants were ages 18-25, 33% were ages 26-35, 17% were ages 36-45, 10% were ages 46-55, and 3% were over age 55. Approximately 20% of the participants were African-American or Black, 5% were Latino, 50% were White, and 17% were Asian. Fifty-seven percent of the participants were women.
The two-page questionnaire took approximately three to five minutes to complete. In it, we asked participants to read a brief vignette describing a crime, and then to respond to the questions that followed. (85) Participants were advised that their answers were completely anonymous, and that we would not be asking for their names.
2. Materials and Design
We tested participants' reactions to one of two (86) different crimes: each participant read either a burglary (208 participants) or a robbery (94 participants) vignette. (87) Within each crime, we randomly assigned participants to read one of three different victim impact statements. (88) We did not tell the participants that there were three different victim statements. Rather, each participant read and made judgments about only one version of the crime. (89) Participants receiving the Severe Emotional Injury statement read that, as a result of the crime, the victim is now feeling afraid, vulnerable, depressed, is having problems sleeping, and cannot stop thinking about the crime. In the Mild Emotional Injury statement, participants read that the victim was angry when the crime first happened, but now has returned to her normal activities and no longer thinks too much about the crime. Finally, in the Control condition, participants received no information about how the victim was coping with the crime.
The Control condition allows us to compare the average sentence judgment when no victim impact information is present to the average sentence judgment when victim impact evidence is present. If sentencing judgments are, indeed, influenced by the adventitious harm presented in victim impact statements, we would expect differences in sentence judgments across all three groups. That is, compared to the Control condition, we expect that evidence of a victim's Severe Emotional Injury would lead participants to render particularly punitive sentences, whereas evidence of a victim's relatively Mild Emotional Injury would yield comparatively lighter sentences. The degree of adventitious harm would be a less reliable indicator of the severity of punishment judgments if, compared to the Control condition, sentence judgments increase when any victim impact information is presented. This result would suggest that mere victim-focus, rather than extent of harm, is the key variable in the relationship between victim impact testimony and sentence judgments.
After reading the crime vignette (and victim statement, if any), participants were asked to indicate what prison sentence, if any, the defendant should receive for the crime. Participants responded on a scale that ranged from "Probation (No Prison)" on one end to "18 years or more" on the other end. (90) Participants then answered a number of other questions regarding their opinions as to the seriousness of the crime and the extent of injury to the victim, as well as demographic questions about themselves.
We first examined Whether the victim impact statement in the Severe Emotional Injury version of the questionnaire was, in fact, perceived by the participants as describing a more severe harm than that represented in the Mild Emotional Injury version of the questionnaire. Indeed, this was the case for both the burglary and the robbery vignette. Participants were asked to rate, in their own opinion, how upset the victim was on a scale from 1 to 7. Participants who read the victim impact statement describing Severe Emotional Injury rated the victim as significantly (91) more upset, on average, than participants who read the Mild Emotional Injury statement, for both burglary (92) and robbery. (93) On this question, the victim in the Control version of the questionnaire was rated as being, on average, between the Severe and Mild responses for both crimes. (94) Thus, we successfully manipulated the severity of emotional harm. (95)
We then examined our main question of interest, namely whether the emotional severity of victim impact evidence influenced the severity of sentences imposed on defendants. Our results indeed demonstrate such an effect of victim impact evidence on severity of sentence. In the burglary vignette, participants chose a higher prison term, on average, when the victim impact evidence described Severe Emotional Injury (4.4 years) than when it described Mild Emotional Injury (2.7 years). (96) In the robbery vignette, a similar pattern emerged: participants assigned more severe sentences, on average, when the victim impact statement expressed Severe Emotional Injury (4.8 years) than when the victim impact statement expressed Mild Emotional Injury (3.1 years). (97)
Participants were also asked about their own emotional reactions to the crime. Specifically, they were asked to indicate on a scale from 1 to 7 the degree to which they felt sympathy, anger, and disgust while reading about the crime. We hypothesized that the expression of severe emotional harms in the victim impact statements would lead mock jurors to experience more intense emotions themselves. The results indicate that this was true at least of participants' feelings of sympathy. For the burglary vignette, participants rated their own feelings of sympathy higher, on average, when the victim impact evidence described Severe Emotional Injury (mean = 5.12) than when it described Mild Emotional Injury (mean = 4.22). (98) For the robbery vignette, the same pattern emerged, although the overall differences were only marginally statistically significant. (99) There were no significant differences in participants' reported experience of feelings of disgust or anger based on the severity of victim impact statements. (100) Finally, participants were asked to rate the seriousness of the crime on a scale from 1 (not serious) to 7 (extremely serious). Although the average rating of crime seriousness was not significantly different based on the severity of victim impact statements, (101) the percentage of participants rating the crime as very serious (indicated by a rating of 6 or 7) was significantly greater when victim impact statements were Severe (15%) than when they were Mild (3.5%). (102)
These data show that as the victim describes increasingly severe emotional harm, the sentence imposed on the defendant likewise increases. (103) Laypersons (and perhaps judges as well) (104) are more punitive when the outcome of the crime reflects greater emotional harm to the victim. Even when all other circumstances surrounding the crime and the defendant are held constant, the punishment is more severe when the victim is psychologically less able to deal with the crime in its aftermath. In addition, the defendant who had the "luck" of committing his crime on a victim who coped well benefited from a shorter sentence, as compared with the Control vignette. Thus, punishment tracked the adventitious harm described in the different vignettes.
But does the fact that people take this adventitious harm into account when making punishment judgments mean that we should give these intuitions credence? It is a separate question whether and to what extent the emotional reaction of the victim should influence punishment judgments. The answer to this question depends, in part, on the psychology of punishment. If the psychological mechanism by which emotional victim impact influences punishment judgments serves a useful social function, or is recognized as socially desirable, then this conclusion might constitute a reason to defer to popular intuitions. On the other hand, if the jury's consideration of emotional impact on victims is recognized as socially undesirable, there may be a reason to limit or even eliminate the sentencing authority's reliance on such evidence. In the next Part, we discuss several social psychological mechanisms underlying the influence of victim impact evidence on punishment judgments. For each social psychological mechanism we discuss, we argue that there is cause for concern that victim impact statements produce punishment judgments that are less reliable than they would be in the absence of such evidence. In the Conclusion of this Article, we make recommendations for limiting the use of victim impact statements based on the analysis in Part IV.
SOCIAL PSYCHOLOGICAL PROCESSES THAT ACCOUNT FOR THE INFLUENCE Of VICTIM IMPACT EVIDENCE ON PUNISHMENT JUDGMENTS
A. Outcome Bias: Emotional Harm as a Heuristic for Culpability
Evidence about the emotional impact of a crime on the victim or the victim's survivors is by nature subjective and more difficult to quantify than objective physical or financial harms. Such evidence is also by nature unforeseeable, because no one can predict the emotional impact that a crime will have on the victim and others. Despite its subjective and unforeseeable nature, people support the use of victim impact statements. (105) Moreover, as shown in Part III, victim impact statements do influence decisions about punishment. In this subpart, we argue that in cases in which culpability and punishment judgments are informed by adventitious results (such as emotional harms described in victim impact statements), those judgments are likely to be unduly influenced by knowledge of the adventitious outcome. Before addressing the argument that victim impact statements are a source of outcome bias, we first address adventitious harm generally, and why victim impact statements are an especially unreliable species of adventitious harm.
Recall that the criminal law tracks popular intuitions about the role of harm, even when harm is adventitious. For example, the successful assassin is punished more harshly than the unsuccessful assassin. (106) From the philosophical perspective of retribution, it is not immediately obvious why we would draw such a distinction, given that, in both instances, the offender performed the very same act with the very same mental state. And because we know that, as an empirical matter, people are retributivists who are motivated to base punishment-judgments on moral culpability, (107) it is not clear why people prefer harsher punishments solely on the basis of adventitious outcomes. That is, why do commonsense judgments of blame and punishment take into account harm that is produced by chance factors? This question is especially puzzling with respect to subjective, emotional harms of the type witnessed in victim impact statements. This question is best illustrated with an example: (108)
Case A: A driver recklessly drives through a red light at a busy intersection. There are numerous pedestrians present. By chance, no pedestrians happen to be in the car's path, and no one is hurt.
As Adam Smith observed with regard to the stone-throwing example, most people would be shocked if this driver received a severe punishment, such as a lengthy prison term. (109) In essence, people reason that because there was no harm, there is no foul.
Case B: The same driver driving the same car through the same intersection commits the exact same act (drives through the red light) with the exact same reckless state of mind. As before, there are numerous pedestrians present. But in this case, the driver hits and kills four pedestrians who happen to be crossing the street.
In Case B, most people would not be particularly shocked to hear that the driver received a lengthy prison term. For the purposes of this example, we can assume that everything about the two situations is exactly the same except for the outcome, which is based on chance factors. When asked to evaluate moral blame and assess punishment, we severely blame and punish B and do not blame or punish A. Why? (110)
One possible psychological explanation for our strong intuitive sense that harm matters (even if adventitious) is that the outcome of an act serves as a heuristic for the actor's effort and other factors that are difficult to ascertain. (111) The social world is filled with uncertainties, and it is impossible to recognize every relevant detail of a complex social event, such as a criminal homicide. In the example above, we do not know what the driver was thinking at the time of the reckless act, and why he ended up driving through the red light. Did he take his eyes off the road? If so, why? Or did he misjudge the distance between his car and the intersection and try to make it through the yellow light? Or did he know that the light would be red when he went through it but sped up and hoped for the best? Given the well-documented limitations of first-person report, (112) the driver's exact motivations and mental state cannot be determined with any certainty. We cannot know much more than this: a car went through a red light, and people either were or were not hurt. We can certainly infer from this evidence important relevant factors such as the driver's state of mind. But such inferences do not provide reliability. In conditions filled with this much uncertainty, harm can be an indicator for what really occurred. In this sense, outcome is a proxy for the defendant's state of mind, motive, and ultimately, his culpability and deservingness of punishment. Harm is a heuristic that helps us interpret what happened in the absence of better information. (113)
Empirical support exists for the theory that harm serves as a heuristic by helping us fill in the blanks regarding unknown factors about the crime. In a study that tested memory for eyewitness events, Elizabeth Loftus found that eyewitness memory of the details of a videotaped car accident varied with the severity of the crash (as characterized by lawyers' leading questions). (114) For example, eyewitnesses estimated that the car involved in the accident was traveling at faster speeds, and they were more likely to report having seen broken glass (although there was none) if the car was described as having "smashed" the other car (as opposed to having "contacted" it). (115) Generally, when harm seems severe, we infer that the causal factors associated with the harm are also severe. In the context of victim impact statements, Edith Greene and her colleagues found that when murder victims were presented as more respectable, mock jurors perceived the victim's survivors as suffering more, and the crime in question as more severe, than when murder victims were presented as less respectable. (116) These studies illustrate that the severity of the harm shapes inferences about the severity of the causes of that harm.
Professor Larry Alexander has suggested that despite our intuitions to the contrary, the criminal law should focus solely on the culpable act (actus reus + mens rea), to the complete exclusion of social harm. (117) It is not clear, however, that a sentencer can mentally separate the culpable act from the social harm. Upon learning the extent of the harm, the sentencer uses this information, consciously or unconsciously, as a basis for judgment. The assessment of the culpable act (i.e., how responsible was the defendant for what happened) is driven by the extent and nature of the social harm. Consequently, there exists a "severity-dependent attribution" of responsibility--the more severe the outcome, the more responsible the defendant seems, even if all other factors are held constant. (118) This concept is best illustrated by a study discussed above, in which respondents rated the responsibility of a bank robber who had aimed and fired his gun at a bank teller but accidentally hit a customer. (119) In one version of the story, the harm to the customer was quite mild (the bullet only grazed him), while in the other version, the harm was more severe (the bullet paralyzed him). (120) Participants rated the defendant as more responsible for the incident when the customer suffered paralysis than when he only suffered minor injuries. (121) It is therefore problematic to say that the criminal law should worry only about the culpability of the act, because sentencing decisions made by applying criminal law principles are unavoidably colored by the severity of the harm.
The fact that severity of harm molds our inferences about the cause of the harm might explain why many people felt that a severe sentence was well deserved in the recent case of the Boston father who killed another father in a fist-fight at their kids' hockey game. (122) The defendant, Thomas Junta, received a sentence of six to ten years of imprisonment, a punishment that many considered to be harsh, given that the degree of criminal homicide for which he was convicted was involuntary manslaughter. (123) But at the same time, the public largely regarded the sentence as just. Why? The harm-as-heuristic theory suggests that it is because of the severity of the harm that resulted from the father's crime. Fist-fighting is conduct that normally does not lead to great harm, and for this reason does not often lead to harsh punishment. According to the harm-as-heuristic model, popular intuition demands harsh punishment for Junta because the result of his crime--death--serves as a rule of thumb that allows us to make inferences about the culpability of the act. Specifically, the severe harm to the victim gives rise to the inference that the act was truly dangerous and reckless. Then, based on the heightened culpability of the act, we punish harshly. Accordingly, we would punish another defendant involved in a fist-fight less harshly if the resulting harm were much less, because the recklessness of the act would be uncertain without firm evidence of the harm. Though operating as a heuristic, people view harm as much more than a rough estimate: harm provides strong evidence of the severity of the crime.
But how reliable is harm as evidence of culpability? At this point, it is useful to examine whether some types of harm might be more reliable indicators of the general culpability of a defendant than others. In particular, victim impact statements describe a type of harm that is particularly unreliable as a heuristic regarding unknown factors related to culpability. To understand why, let us first consider physical harm, which is a standard feature of the prosecution's case during the guilt phase of a trial. In most cases, the extent of physical harm that a victim suffers is a relatively reliable indicator of factors related to culpability, as illustrated by the Junta case discussed above. Few morphologically important individual differences are likely to exist between persons that would leave one person dead and the next person unscathed following the very same act. That is, regardless of who the victim is, the extent of bodily injury reveals a great deal about the act that led to the injury. (124) However, unlike physical harm, emotional harm is a particularly unreliable heuristic, became individual emotional responses to a crime vary widely among individuals. In addition to the wide range of potential emotional reactions, how effectively victims and survivors express their experiences in court is likely to vary broadly. (125) That is, unlike physical injury, the severity of emotional injury permits few reliable inferences about the act that led to the injury, because victims vary widely on how they respond and express themselves emotionally to the same crime. Thus, although it may be intuitively compelling to punish offenders more harshly when the resultant emotional harm is more severe, this intuition probably derives from the general proposition that harm is a reliable indicator of defendant culpability. In the case of emotional harm, however, the severity of such harm might not be a reliable indicator of defendant culpability. In applying the general principle that harm is a good indication of culpability, we fail to discount for the unreliability of emotional harm.
Evidence suggests that people are susceptible to outcome bias. That is, people evaluate a decision or other action more favorably when the outcome is positive than when it is negative, to a greater extent than is warranted by the available information. (126) For example, when people were asked to judge medical decisions (using the same information available to the doctor at the time), they blamed the doctor more harshly and labeled her less competent when the outcome was unfavorable than when it was favorable. (127) In essence, people take into account information about the outcome of an event when judging the person who caused it, even when the outcome is not a reliable basis for judging. As we have argued, compared to other types of harm resulting from a crime, such as physical harm, the emotional harm contained in victim impact statements is not a reliable proxy for judging relevant but unknown circumstances of the crime, such as the defendant's state of mind. Yet, we are drawn to outcome information even when we are aware of its unreliability. In the aforementioned study in which participants judged a doctor's medical decision, most participants recognized that they should not consider outcomes in their evaluations. (128) Despite knowing that outcome information was not a reliable tool in evaluating the doctor's decision, they used outcome information nonetheless. (129) Similarly, jurors hearing victim impact statements might recognize that the information is an unreliable heuristic for judging the unknown circumstances surrounding the crime. Because of people's general susceptibility to using outcome information in judging behavior, however, jurors are likely to be unduly influenced by victim impact statements.
B. Decision-Maker Emotion
Up to this point we have focused on cognitive social psychological processes that help to explain why people generally support allowing the sentencer to consider victim emotional harm, as well as how evidence of such harm influences decisions about punishment. To focus exclusively on cognitive processes would be to ignore one of the main concerns expressed in the Payne dissenting opinions and in subsequent scholarly commentary: victim impact statements might arouse intense emotions that overwhelm the capital sentencing decision process. (130) These concerns explicitly reflect an assumption that the presence of emotion in legal decision making is anathema to careful reasoning. In fact, although the precise relationship between emotional and cognitive processes is both complex and controversial, (131) psychological research has uncovered several ways that emotions can affect decision making.
First, anger is a punitive emotion: people who are angry in reaction to hearing about a crime are motivated to call for greater punishment for that crime. (132) Indeed, in a very general sense, the experience of anger is accompanied by a desire to attack, (133) an instinct that is easily translated into a desire for punishment. Thus, to the extent that the sentencer's anger toward the defendant is evoked or exacerbated by victim impact statements, we would expect sentencers to prefer harsher punishment judgments.
Second, the very demonstration of negative emotions can alter people's explanations for why an event happened. For example, people who were induced in an experimental setting to feel anger were more likely to blame individual persons for negative outcomes (e.g., an arsonist is to blame for a house fire) than they were to blame impersonal forces, such as bad luck (e.g., an electrical shortage is to blame for a house fire). (134) Anger has this effect because the emotion tends to come about, in the first instance, through the recognition that another person has done something to harm us, or has harmed someone with whom we sympathize. (135) Once this focus on individual accountability emerges, it is more salient and readily available when we assess other situations. (136) For example, people who were induced in an experimental setting to feel anger were later asked to assess a completely unrelated tort case. Angry people awarded more damages to the tort plaintiff and judged the defendant to be more deserving of punishment than people who were not angry at the time, even though the anger was unrelated to the tort case. (137)
One explanation for the anger-punishment relationship is that exposure to intense emotional suffering heightens decision makers' negative affect and consequently activates "blame-validation processing." (138) In this state, jurors or other decision makers look for ways to hold an offender responsible, even for unforeseeable outcomes. (139) In response to these negative feelings, people rationalize greater blame for an event--and presumably, concomitant levels of punishment--by altering their evidentiary standards. (140) According to the blame-validation model, evidence that a victim or a family member of a victim has suffered severe emotional injury is not only likely to arouse a decision maker's desire to blame, but will also produce a cognitive search to validate that blame. (141) In short, the anger aroused by hearing victim impact statements often induces decision makers to engage in a search for evidence to support more blame and punishment. (142) Accordingly, despite the fact that the Payne Court reasoned that victim impact statements help restore the balance between defendant and victim, (143) there is reason to suppose that victim impact statements further tip the scales against a defendant.
There is additional evidence to suggest that victim impact statements generate negative affect within the decision maker, and therefore lead to a higher likelihood that the decision maker will search for reasons to blame and punish. Recent research on the "moral emotions" identifies contempt, anger, and disgust as three key responses to wrongdoing by others. (144) Victim impact statements plausibly activate each of these emotions. As previously mentioned, anger might arise through the recognition that someone else is responsible for harm that has befallen another or ourselves. (145) One scholar characterizes this anger as arising through "violations of autonomy." (146) In other words, anger occurs in response to perceived violations of a person's rights or to his well-being. (147) Sentencers naturally perceive the crime itself as a violation of the victim's autonomy, and this perception is likely to elicit anger. By focusing on the continuing emotional impact of the crime, victim impact statements indicate that the violation of autonomy inherent in the crime continues, but in a different form (i.e., by impairing the well-being of the victim's family and friends). This perception of a continuing violation of autonomy is likely to further increase anger in decision makers, and is likely to be associated with more severe punishment.
Apart from anger, victim impact statements are likely to elicit other negative emotions in the sentencer. Contempt--a weaker emotion than anger, but still linked with negative arousal--often arises in response to violations of the ethics of community. (148) Harms to community include violations generally considered less serious in our culture than crimes, such as disrespect for authority or for institutions. (149) The empirically established connection between contempt and community harm is particularly critical for the role of victim impact statements in capital trials, given that in some jurisdictions courts routinely address evidence about how the victim's death has affected the community. (150) Whether contempt arising from community harm influences punishment judgments is still an open question that has not been resolved empirically. But it is worth noting that victim impact statements often focus attention not only on the harm to the individual victim and individual family members and friends, but also on the way in which the crime has affected a larger social group. (151) In short, there are several ways that negative emotions directed at the defendant can lead to more blame, a change in evidentiary standards, and ultimately more punishment.
Finally, apart from engendering anger toward the defendant, another way that victim impact statements might influence sentencing decisions is through the decision maker's sympathy for the victim. People who feel sympathy for someone's plight will likely be motivated to help that person. (152) Therefore, to the extent that decision makers believe that assigning harsher punishment to an offender will help the victim, and will relieve the distress of family members and friends, the victim's or family's story that elicits greater sympathy should also elicit more severe punishment. Similarly, to the extent that decision makers perceive punishment as "helpful" to the victim's family by putting themselves in the family's position and asking what they would want in that situation, decision makers are likely to assume that the family would want the harshest possible punishment for the defendant. (153)
A particularly intriguing aspect of the role of sympathy in punishment judgments is that the emotional intensity of witnesses giving victim impact testimony is perceived by observers as a cry for help or relief. The decision maker has at her disposal the power to punish the offender and the power to decide the severity of the punishment. In this way, the emotion expressed by the victim becomes a covert communication about the victim's, or the victim's family's, desire for retribution. Consequently, this expression obviates the Payne Court's prohibition of victim opinion testimony regarding appropriate punishment. (154) Through the implicit evocation of sympathy, the victim can communicate his or her wishes for severe punishment.
Our data support a link between sympathy for a victim and levels of punishment. Interestingly, the results highlight the extent to which chance factors can affect punishment decisions, for it was the failure on the part of the victim expressing mild emotional harm to elicit sympathy among participants in our study that was associated with less punishment. Subjects exposed to a victim's mild emotional harm reported less sympathy for the victim--who was, after all, coping quite well--and recommended less punishment for the offender. (155) This resilient victim was perceived as not needing the help that could be offered through punishment. Accordingly, subjects believed that the defendant deserved less punishment.
Now imagine the reverse situation in the context of a capital sentencing hearing: intensely emotional victim impact statements provided by family, friends, and neighbors, describing the exemplary character of the victim who is now gone forever and the painful emotional toll on the survivors, evoke sympathy to such an extent that it becomes more probable that the jury will perceive death as the most appropriate punishment. In either event, the emotional nature of victim impact statements makes real the very danger that the Court worried about in Booth: "defendants whose victims were assets to their community are more deserving of punishment than those whose victims are perceived to be less worthy." (156) In overruling Booth, the Payne majority dismissed this danger by declaring that "victim impact evidence is not offered to encourage comparative judgments of this kind." (157) However, our analysis of recent scholarly work on the influence of emotion on blame and punishment judgments, together with our own empirical results regarding the role of sympathy on punishment judgments, suggest that this danger cannot be so easily dismissed.
C. Victim Impact Statements as Clinical Interviews: More Information Is Not Always Better
We have discussed several psychological processes that are likely to come into play during jurors' consideration of victim impact statements, and have set forth why each of these processes, standing alone, makes it likely that the introduction of victim impact statements will affect punishment decisions regarding severity. We now discuss a final reason for concern: the counterintuitive, yet robust, empirical finding that more information is not always better, and that consideration of certain types of information can lead to decisions that are less reliable than decisions made in the absence of the additional information. We begin by returning to the Court's reasoning in Payne.
In Payne, the Court's justification for allowing victim impact statements was grounded in part on "balancing" the need for information about the victim against the defendant's right to present unfettered mitigating information. (158) It would be unfair, the Court reasoned, to have a one-sided presentation that included everything the defendant could muster in mitigation, while including nothing about the victim, because the State would be deprived of "the full moral force of its evidence and [it would] prevent the jury from having before it all the information necessary to determine the proper punishment for a first-degree murder." (159)
Thus, the concern for balancing and avoiding a one-sided presentation assumes that victim impact statements contain important information that will aid the jury in making their decision. Implicit here is the assumption that "more is better"--i.e., more information makes for a better decision. However, one problem with this argument is that more information does not, in fact, always result in a better decision. In many settings in which it is possible to evaluate the quality of decisions, the addition of certain types of information leads to a less reliable, lower-quality decision. (160) This is because human decision makers tend to make systematic errors in the way they combine pieces of information. Specifically, decision makers tend to overweight certain types of information and underweight others. (161) Under conditions of improper weighting, decision makers often are better off not considering the extra information that tends to be overweighted. (162) Therefore, more is not always better. Unfortunately, as humans we are disposed to assume that more is better, and that if information is nondiagnostic or non-informative, we will know it when we see it. We are inclined to think that once we have all the information available, we will use our reason to consider all of the pieces of information. (163)
Certain classes or categories of information are especially prone to being overweighted in decision making, and the use of such unreliable information can lead to worse decisions. An example of this phenomenon is the use of interviews by gatekeepers. (164) For example, when managers in a business organization or admissions officers in an educational institution are given the opportunity to conduct a face-to-face interview with a candidate, they make use of the information gleaned from the interview to make their hiring or admission decisions. (165) Unfortunately, decisions that take into account the interviews are systematically worse than decisions made using a simple model that includes just a few factors that reliably predict success. (166) In other words, a gatekeeper who makes a decision based in part on an interview makes a less reliable decision than a gatekeeper who uses a simple statistical prediction model (such as admitting the applicants with the highest test scores) and no interview. People incorrectly assume that they can glean a much deeper understanding of a candidate in a forty-five minute interview than they can from a few simple predictive factors such as test scores.
Another example is the problem of predicting parole violations. (167) A statistical model of parole success that includes the offender's age at first arrest, number of convictions, and number of prison violations predicts with fairly poor success which parolees are likely to violate parole (correlation coefficient of about .30). (168) However, the prediction of experienced parole officers who use an interview to predict parole success is much worse (correlation coefficient of .06). (169) In other words, the decision makers would be better off ignoring all information gleaned from the interview, and simply combining the three aforementioned factors using a linear model.
The use of victim impact statements may be problematic for some of the same reasons, although it would be difficult to prove this assertion because, unlike decisions that involve predictions about future events, there is no straightforward way of evaluating the quality of jury decisions. Yet, there is still a danger that rather than serving a balancing function, victim impact statements might erode the judgment of decision makers by causing them to overweight factors contained in such evidence. This overweighting of information in victim impact statements is especially worrisome in light of the psychological mechanisms implicated in the relationship between harm and punishment judgments discussed earlier. (170) First, information about emotional harm is used as a proxy for determining the severity of causal factors leading to that harm, even when the relevant causal factors are accidental. (171) Second, decision makers who hear accounts of emotional harm are likely to experience anger and other negative affective states, increasing the likelihood that they will blame and severely punish the offender. (172) Finally, receiving information about emotional harm triggers a search for information by decision makers that will support the judgment about the severity of the harm. (173) In sum, there is a danger that particularly severe emotional harm suffered by the victim or the murder victim's family members will lead to an exaggerated view of the defendant's culpability, because decision makers will use the cue of severe harm to systematically overweight unknown but important factors about the crime. This systematic overweighting is prone to occur because, as we argue above, emotional harm, unlike physical harm, varies so much across individuals that it is not a particularly reliable indicator of relevant factors about the crime. (174)
Even in the absence of wide variations in emotional harms across individual victims, there is another reason why victim impact statements might make the jury's decision more difficult. Susan Bandes has argued that victim impact statements are stories that should not be told, at least not in the context of capital sentencing, because they interfere with the jury's ability to hear the defendant's story. (175) In a similar vein, Angela Harris has argued that victim impact statements interfere with jurors' ability to reflect on their own emotional reactions to the defendant in a self-critical manner. (176) Both Bandes and Harris contend that victim impact statements not only evoke sympathy, pity, and compassion for the victim, but they also evoke emotions directed toward the defendant, including hatred, racial animus, vindictiveness, and vengeance. (177) According to this perspective, rather than simply making the victim more human in the eyes of the jury, victim impact statements interfere with the jury's ability to empathize with the defendant and comprehend his humanity. In this sense too, then, more information is not always helpful, because the additional information contained in victim impact statements prevents the jury from properly considering information it possesses about the defendant.
Finally, victim impact statements might erode the jury's judgment by lulling jurors into a false sense of understanding. If an event, especially an event with a terrible outcome, has a causal explanation that is comprised of random factors or multiple situational factors that are only tenuously related to human agency, we have trouble accepting such an explanation. Instead, we prefer to hold the simpler and more satisfying view that bad things are caused by bad people whom we can blame and hold accountable. For example, after TWA flight 800 exploded over Long Island, New York, in 1996, rumors quickly spread that a missile was the cause of the crash, leading to speculation of a U.S. government cover-up. (178) Meanwhile, the NTSB had concluded that the most likely explanation for the crash was an electrical short circuit that caused the airplane's fuel tank to explode. (179) Even today, however, rumors of a government conspiracy abound, precisely because the fuel tank explosion explanation is deeply unsatisfying in that there is no human agent to blame.
Returning to the context of the criminal trial, if decision makers use victim impact statements as a heuristic to infer unknown factors of a crime--especially those regarding the defendant's mental state, motivation, and wickedness--the severity of emotional harm will encourage decision makers to infer the worst about the circumstances surrounding the crime. In allowing the use of victim impact statements, courts put decision makers in a position analogous to the TWA Flight 800 conspiracy theorists by promoting salient causal factors involving human agency, and downplaying the random, situational causal factors that contribute to victim suffering. Victim impact statements provide decision makers with a satisfying sense of understanding, but at the expense of imposing a decision procedure that is more likely to push the decision maker toward an unduly severe punishment judgment.
CONCLUSION: REDUCING ARBITRARINESS IN CAPITAL SENTENCING JUDGMENTS
The empirical evidence presented in this Article suggests that judgments about criminal punishment are influenced by the content of victim impact statements. We have argued that the emotional harm to victims and survivors presented in victim impact statements serves as a heuristic for inferring facts that are essentially unknowable. At the same time, we have reason to suppose that emotional harm to victims is not a particularly reliable heuristic for making such inferences. The emotional nature of victim impact statements increases the possibility that judges and juries will allow their own emotions to influence their sentencing decisions. Exposure to victims' emotional harms serves to conveniently confirm a judgment that the decision maker has made earlier in the case, or that he or she will make later in the case. Finally, victim impact statements might be another example of the "more-is-worse" phenomenon. That is, the complexity of the information contained in victim impact statements makes it unlikely that this information will be combined and utilized in a way that reliably assists the sentencing decision.
The reality of how people reason about harm and punishment suggests that the use of victim impact statements is problematic. Yet, this assertion is difficult to prove conclusively because, unlike decisions that involve a verifiably correct answer, there is no straightforward way of evaluating the quality of jury decisions about punishment. The dangers identified in this Article regarding the undue influence of victim impact statements on punishment judgments counsel against their use in criminal trials, and especially in death penalty hearings. At the same time, however, the political reality is that the use of victim impact statements in criminal trials is not likely to disappear entirely, at least not in the foreseeable future. The victims' rights movement is already politically powerful and continues to gain momentum. In addition, there is widespread popular support for the notion that the victim's voice ought to be heard at the criminal trial. Proponents of victims' rights explicitly endorse the idea that juries should hear victim impact statements in capital cases because it provides them with information about the harm the defendant has caused. (180)
Given this political reality, the most realistic way of limiting the arbitrariness produced by victim impact statements is to limit their scope. A noteworthy feature of Payne is that, aside from delineating permissible substantive areas of the victim's character and the survivors' emotional harm, (181) the Court placed virtually no limits on the permissible scope of victim impact statements in capital sentencing hearings. The result has been that, in the twelve years since the Court decided Payne, prosecutors across the country have been presenting an array of testimony, exhibits, and argument in capital sentencing hearings under the rubric of victim impact statements. (182) By narrowing the scope of admissible information contained in victim impact statements, courts can reduce the inherent arbitrariness of their effects. (183)
VIGNETTES USED IN THE STUDY
1. Burglary Vignette
Severe Emotional Injury:
Sharon Martin's apartment was burglarized while she was away. She had gone to visit her parents for the weekend. When she returned, she walked into her apartment and found that someone had come in and ransacked it. Her belongings were thrown all over the floor. Her jewelry and her computer were stolen. Several weeks later, a suspect was arrested when he was caught trying to break into another apartment in the neighborhood. The defendant confessed to the burglary of Sharon Martin's apartment, and eventually pled guilty in court. At the sentencing hearing, the judge hears the defendant's account of the crime, and hears from Sharon Martin. She says that the burglary has made her feel very afraid, vulnerable and depressed. She stayed home from work for two days after the burglary. Now she is back at work, but she still has problems sleeping. She finds herself worrying often that something similar will happen to her again, and she can't stop thinking about the fact that she will probably never get her jewelry or her computer back. The defendant has a criminal record that includes several burglaries and thefts.
Mild Emotional Injury:
Sharon Martin's apartment was burglarized while she was away. She had gone to visit her parents for the weekend. When she returned, she walked into her apartment and found that someone had come in and ransacked it. Her belongings were thrown all over the floor. Her jewelry and her computer were stolen. Several weeks later, a suspect was arrested when he was caught trying to break into another apartment in the neighborhood. The defendant confessed to the burglary of Sharon Martin's apartment, and eventually pled guilty in court. At the sentencing hearing, the judge hears the defendant's account of the crime, and hears from Sharon Martin. She says that when the burglary first happened, it made her feel angry. But after several days her anger faded, and she went about normal activities without thinking much about what happened. She feels like these types of things are bound to happen to people living in a big city, and she doesn't let it bother her, even though she knows she will probably never get her jewelry or computer back. The defendant has a criminal record that includes several burglaries and thefts.
Sharon Martin's apartment was burglarized while she was away. She had gone to visit her parents for the weekend. When she returned, she walked into her apartment and found that someone had come in and ransacked it. Her belongings were thrown all over the floor. Her jewelry and her computer were stolen. Several weeks later, a suspect was arrested when he was caught trying to break into another apartment in the neighborhood. The defendant confessed to the burglary of Sharon Martin's apartment, and eventually pled guilty in court. At the sentencing hearing, the judge hears the defendant's account of the crime, and hears from Sharon Martin. She says that she knows she will probably never get her jewelry or computer back. The defendant has a criminal record that includes several burglaries and thefts.
2. Robbery Vignette
Severe Emotional Injury:
Renee Thompson was robbed of her purse while walking down the street. She had just come out of a Walgreen's store a few minutes before when she felt someone yank her purse from behind and knock her to the ground. The purse contained $120 cash, as well as credit cards, a cellular phone, and other personal items. She suffered a fractured wrist and scrapes and bruises. The police arrested the defendant the next day after a witness came forward. They found Ms. Thompson's credit cards in the defendant's pocket. The defendant eventually pled guilty to robbery. At the sentencing hearing, the judge hears the defendant's account of the crime, and hears from Renee Thompson. She says that the robbery has made her feel very afraid, vulnerable, and depressed. She stayed home from work for two weeks after the robbery. Now she is back at work, but she still has problems sleeping. Before the robbery, she used to have an active social life with many friends; but ever since the robbery she rarely goes out. She finds herself worrying often that something similar will happen to her again, and she can't stop thinking about what happened. The defendant has prior crimes on his record, including two armed robberies where he displayed a gun.
Mild Emotional Injury:
Renee Thompson was robbed of her purse while walking down the street. She had just come out of a Walgreen's store a few minutes before when she felt someone yank her purse from behind and knock her to the ground. The purse contained $120 cash, as well as credit cards, a cellular phone, and other personal items. She suffered a fractured wrist and scrapes and bruises. The police arrested the defendant the next day after a witness came forward. They found Ms. Thompson's credit cards in the defendant's pocket. The defendant eventually pled guilty to robbery. At the sentencing hearing, the judge hears the defendant's account of the crime, and hears from Renee Thompson. She says that when the robbery first happened, it made her feel angry. But she went back to work right away, and after several days her anger faded, and she went about her normal activities without thinking much about what happened. She still has an active social life with many friends. She feels like these types of things are bound to happen to people living in a big city, and she doesn't let it bother her. The defendant has prior crimes on his record, including two armed robberies where he displayed a gun.
Renee Thompson was robbed of her purse while walking down the street. She had just come out of a Walgreen's store a few minutes before when she felt someone yank her purse from behind and knock her to the ground. The purse contained $120 cash, as well as credit cards, a cellular phone, and other personal items. She suffered a fractured wrist and scrapes and bruises. The police arrested the defendant the next day after a witness came forward. They found Ms. Thompson's credit cards in the defendant's pocket. The defendant eventually pled guilty to robbery. At the sentencing hearing, the judge hears the defendant's account of the crime, and hears from Renee Thompson. The defendant has prior crimes on his record, including two armed robberies where he displayed a gun.
We are grateful to the American Bar Foundation for financial support for this research. We would like to thank Shaft Diamond for helpful comments on an earlier version of this Article. We are also grateful to Amanda Dykema-Engblade, Esther Choi, and Cristina Lane for excellent research assistance.
(1) ADAM SMITH, THE THEORY OF MORAL SENTIMENTS 152 (London, Harrison & Sons 1853) (1759).
(2) See id.
(3) Id. at 148-49.
(4) Id. at 149.
(5) In California, for example, the sentence for an attempt is one-half the sentence otherwise applicable for the completed crime. CAL. PENAL CODE [section] 664 (West 1999).
(6) In addition to these empirical questions about how variations in harm influence people's attitudes about punishment, there are also normative questions about how such variations in harm ought to affect punishment. For the purposes of this Article, we focus on the former question.
(7) 501 U.S. 808 (1991).
(8) See, e.g., Susan Bandes, Empathy, Narrative, and Victim Impact Statements, 63 U. CHI. L. REV. 361, 392-93 (1996) (noting that victim impact statements are stories that should not be told in the context of capital sentencing because they overwhelm the jury and interfere with its ability to empathize with the defendant); Angela P. Harris, The Jurisprudence of Victimhood, 1991 SUP. CT. REV. 77, 96-101 (noting that the jury's focus on the victim during the sentencing phase will lead to stereotyping and insufficient weight to mitigating factors); Wayne A. Logan, Through the Past Darkly: A Survey of the Uses and Abuses of Victim Impact Evidence in Capital Trials, 41 ARIZ. L. REV. 143, 191-92 (1999) (suggesting that limits on the use of victim impact statements are necessary to reduce the resulting arbitrariness in capital sentencing); Niru Shanker, Getting a Grip on Payne and Restricting the Influence of Victim Impact Statements in Capital Sentencing: The Timothy McVeigh Case and Various State Approaches Compared, 26 HASTINGS CONST. L.Q. 711, 740 (1999) (concluding that the prejudicial effect of victim impact statements must be restricted).
(9) 408 U.S. 238 (1972).
(10) 428 U.S. 153 (1976).
(11) See Gregg, 428 U.S. at 206-07 (upholding defendant's capital sentence under Georgia's sentencing procedures because those procedures focused the jury's attention on the particularized nature of the defendant and the crime); Furman, 408 U.S. at 256-57 (Douglas, J., concurring) (overturning defendants' capital sentences because the statutes under which they were sentenced were susceptible to arbitrary and capricious application).
(12) See, e.g., Amy K. Phillips, Note, Thou Shalt Not Kill Any Nice People: The Problem of Victim Impact Statements in Capital Sentencing; 35 AM. CRIM. L. REV. 93, 105-06 (1997).
(13) For empirical data on this issue, see Edith Greene et al., Victim Impact Evidence in Capital Cases: Does the Victim's Character Matter?, 28J. APPLIED SOC. PSYCHOL. 145, 154 (1998) (finding that greater victim "respectability" led to judgments that the crime was more serious).
(14) For examples of victim impact statements admitted into evidence but describing harms that are arguably causally remote, see Logan, supra note 8, at 160-65.
(15) See, e.g., JEREMY BENTHAM, AN INTRODUCTION TO THE PRINCIPLES OF MORALS AND LEGISLATION 178-79 (1948); IMMANUEL KANT, THE PHILOSOPHY OF LAW 194-98 (W. Hastie trans., 1887); JEFFRIE G. MURPHY & JEAN HAMPTON, FORGIVENESS AND MERCY 125-26 (1988).
(16) See, e.g., PETER H. ROSSI & RICHARD A. BERK, JUST PUNISHMENTS: FEDERAL GUIDELINES AND PUBLIC VIEWS COMPARED (1997).
(17) See PAUL H. ROBINSON & JOHN M. DARLEY, JUSTICE, LIABILITY, AND BLAME: COMMUNITY VIEWS AND THE CRIMINAL LAW (1995).
(18) See, e.g., Daniel Kahneman et al., Shared Outrage and Erratic Awards: The Psychology of Punitive Damages, 16J. RISK & UNCERTAINTY 49, 62-63 (1998) (finding that jury punishment judgments are affected by juror outrage at the degree of harm suffered by the plaintiff); Cass R. Sunstein et al., Do People Want Optimal Deterrence?, 29 J. LEGAL STUD. 237, 243-44 (2000) (finding that people reject the notion that punishment should vary with probability of detection).
(19) See STANTON WHEELER ET AL., SITTING IN JUDGMENT: THE SENTENCING OF WHITE-COLLAR CRIMINALS 66-68 (1988).
(20) See, e.g., id. at 62-80.
(21) See Sandra Graham et al., An Attributional Analysis of Punishment Goals and Public Reaction to O.J. Simpson, 23 PERSONALITY & SOC. PSYCHOL. BULL. 331, 333, 336 (1997) (describing empirical research that shows retribution to be one of the most important goals of punishment); V. Lee Hamilton & Joseph Sanders, Punishment and the Individual in the United States and Japan, 22 LAW & SOC'Y REV. 301, 318-19 (1988) (same).
(22) See, e.g., John M. Darley et al., Incapacitation and Just Deserts as Motives for Punishment, 24 LAW AND HUM. BEHAV. 659, 660-71 (2000) (presenting empirical results to support the conclusion that when people impose punishment, they are motivated by just deserts considerations).
(23) See Neil Vidmar, Retribution and Revenge, in HANDBOOK OF JUSTICE RESEARCH IN LAW 31, 35 (Joseph Sanders & V. Lee Hamilton eds., 2001) (noting that "retributive justice is concerned with the moral nature of the offense").
(24) See Jonathan Baron & Ilana Ritov, Intuitions About Penalties and Compensation in the Context of Tort Law, 7 J. RISK & UNCERTAINTY 17, 31-32 (1993) (finding that many people do not apply the deterrence rationale when reasoning about compensation and punishment); Sunstein et al., supra note 17, at 239.
(25) See Darley et al., supra note 22, at 666-67.
(26) Id. (contrasting no instruction with retribution and incapacitation strategies); Robert M. McFatter, Sentencing Strategies and Justice: Effects of Punishment Philosophy on Sentencing Decisions, 36 J. PERSONALITY & SOC. PSYCHOL. 1490, 1494, 1498 (1978) (contrasting no instruction with deterrence, rehabilitation, and retribution strategies).
(27) See Mark Warr & Mark Stafford, Public Goals of Punishment and Support for the Death Penalty, 21 J. RES. CRIME & DELINQ. 95, 99-101 (1984).
(28) Jennifer K. Robbennolt, Outcome Severity and Judgments of "Responsibility": A Meta-Analytic Review, 30 J. APPLIED SOC. PSYCHOL. 2575, 2601 (2000).
(29) See Elaine Walster, Assignment of Responsibility for an Accident, 3 J. PERSONALITY & SOC. PSYCHOL. 73, 73-74 (1966).
(30) See id. at 74.
(31) See D. Chimaeze Ugwuegbu & Clyde Hendrick, Personal Causality and Attribution of Responsibility, 2 Soc. BEHAV. & PERSONALITY 76, 78-79 (1974).
(32) See id. at 82-84 (finding that respondents attributed more responsibility to the robber when the consequences of the robber's actions were severe than when they were mild).
(33) Not all scholars agree with the defensive attribution explanation. See, e.g., Marilynn B. Brewer, An Information-Processing Approach to Attribution of Responsibility, 13 J. EXPERIMENTAL SOC. PSYCHOL. 58, 58, 61-62 (1977).
(34) See Robbennolt, supra note 28, at 2601-02.
(35) Harris, supra note 8, at 77.
(36) 482 U.S. 496 (1987), overruled by Payne v. Tennessee, 501 U.S. 808 (1991).
(37) Id. at 505.
(39) See id. at 509.
(40) 501 U.S. 808, 815-16 (1991).
(41) Id. at 828-30.
(42) See id. at 814-16, 830. The Payne decision overturned prior precedent only in capital cases; the sentencing authority's discretion to consider such evidence in noncapital criminal cases was left undisturbed. See id. at 830 n.2.
(43) These examples are loosely based on the victim impact statement presented in Booth, 482 U.S. at 509-12.
(44) PRESIDENT'S TASK FORCE ON VICTIMS OF CRIME, FINAL REPORT 114 (1982) (emphasis omitted).
(45) See Dean G. Kilpatrick et al., The Rights of Crime Victims--Does Legal Protection Make a Difference?, NAT'L. INST. JUST. RES. BRIEF (U.S. Dep't of Just., Wash., D.C.), Dec. 1998, at 1, 1.
(47) Logan, supra note 8, at 150.
(48) See Paul Gewirtz, Victims and Voyeurs at the Criminal Trial, 90 NW. U. L. REV. 863, 868-69 (1996).
(49) South Carolina v. Gathers, 490 U.S. 805, 811-12 (1989) (prohibiting prosecutors' commentary on victim characteristics in capital trials); Booth v. Maryland, 482 U.S. 496, 509 (1987) (prohibiting victim impact testimony in the sentencing phase of capital trials), overruled by Payne, 501 U.S. at 827.
(50) See Payne v. Tennessee, 501 U.S. 808, 825 (1991).
(51) See supra note 5 and accompanying text.
(52) Payne, 501 U.S. at 825-26 (quoting Gathers, 490 U.S. at 821 (O'Connor, J., dissenting)).
(53) See id. at 864 (Stevens, J., dissenting).
(54) See id. at 859-60 (Stevens, J., dissenting).
(55) See id. at 861-62 (Stevens, J., dissenting).
(56) For an excellent summary of these developments, see Logan, supra note 8, at 160-65.
(57) Id. at 164.
(58) The U.S. Supreme Court provided no guidance in Payne, and with a few exceptions, states have provided no limitations of their own. See id. at 151-53.
(59) See Payne, 501 U.S. at 827.
(60) See, e.g., WHEELER ET AL., supra note 19 , at 66-68; Ugwuegbu & Hendrick, supra note 31, at 77-78.
(61) Elizabeth E. Joh argues that victim impact statements allow for the introduction of pain as a sentencing factor, which encourages vengeance. See Elizabth E. Joh, Narrating Pain: The Problem with Victim Impact Statements, 10 S. CAL. INTERDISC. L.J. 17, 19 (2000).
(62) See, e.g., Vidmar, supra note 23, at 41-43.
(63) See Edith Greene, The Many Guises of Victim Impact Evidence and Effects on Jurors' Judgments, 5 PSYCHOL, CRIME & L. 331, 338, 345 (1999); Greene et al., supra note 13, at 149-52.
(64) Greene et al., supra note 13, at 154.
(65) Greene, supra note 63, at 345 (comparing victim as photographer with victim as motorcycle gang member).
(66) See James Luginbuhl & Michael Burkhead, Victim Impact Evidence in a Capital Trial: Encouraging Votes for Death, 20 AM. J. CRIM. JUST. 1 (1995).
(67) Id. at 12-13.
(68) See Bryan Myers & Jack Arbuthnot, The Effects of Victim Impact Evidence on the Verdicts and Sentencing Judgments of Mock Jurors, 29 J. OFFENDER REHABILITATION 95, 107-08 (1999).
(69) See Adelma M. Hills & Donald M. Thomson, Should Victim Impact Influence Sentences? Understanding the Community's Justice Reasoning, 17 BEHAV. SCI. & L. 661,663 (1999).
(70) Hills and Thomson preceded us in examining the adventitious nature of victim impact evidence, although they used the term "fortuitous consequences." Id. at 662.
(71) See id. at 669.
(72) See id. at 664-65.
(73) See Olga Tsoudis & Lynn Smith-Lovin, How Bad Was It? The Effects of Victim and Perpetrator Emotion on Responses to Criminal Court Vignettes, 77 Soc. FORCES 695, 704-05 (1998). For example, in one version of the victim testimony, subjects were presented with a written description that said "I didn't know what to do (weeping, one hand on face)." Id. at 705.
(74) See id. at 709-10.
(75) Id. at 710-11.
(76) See id. at 704-05.
(77) In contrast to the laboratory studies discussed above, studies examining real cases have found that victim impact statements have mixed or no effects on sentencing. See, e.g., Robert C. Davis & Barbara E. Smith, The Effects of Victim Impact Statements on Sentencing Decisions: A Test in an Urban Setting, 11 JUST. Q. 453, 464-65 (1994) (finding no effect); Edna Erez & Pamela Tontodonato, The Effect of Victim Participation in Sentencing on Sentence Outcome, 28 CRIMINOLOGY 451, 467-69 (1990) (finding an effect on sentence outcomes when comparing probation versus incarceration sentencing, but not for sentence length). There are a number of different explanations accounting for the mixed results of these two studies, including the different geographical locations of the courts studied, the distribution of the crimes and victims, the differences in local practices, and the fact that researchers included in the samples cases that had been disposed of through plea bargain. Further, unlike experimental research, potentially confounding factors across cases cannot be controlled.
(78) See Luginbuhl & Burkhead, supra note 66, at 12-13; Myers & Arbuthnot, supra note 68, at 107-08.
(79) See Greene, supra note 63, at 338, 345; Greene et al., supra note 13, at 149-52.
(80) See Hills & Thomson, supra note 69, at 664-65.
(81) See Tsoudis & Smith-Lovin, supra note 73, at 704-05.
(82) In exchange for agreeing to fill out the questionnaire, participants were offered a chocolate chip cookie.
(83) The building houses administrative offices, a bookstore, a cafeteria, and a dormitory, and is also a pickup point for physical rehabilitation patients. The building's lobby is also a popular walk-through area for many area lunch spots.
(84) These participants were almost exclusively medical students. Law students, law professors, and lawyers were not allowed to participate.
(85) Note that although participants essentially acted as mock jurors, limitations on time and the experimental setting did not permit group deliberations as mock juries. Caution must be used in generalizing the individual questionnaire responses in this study to predictions about the behavior of juries, because differences between decisions of individuals and decisions of groups vary in ways that are difficult to predict. See, e.g., Norbert L. Kerr et al., Bias in Judgment: Comparing Individuals and Groups, 103 PSYCHOL. REV. 687 (1996).
(86) The vignettes for the two different crimes are reproduced in the Appendix.
(87) We initially administered the burglary vignette to 208 participants. To ensure that the effects would generalize to other crimes, we then administered the robbery vignette to a smaller sample of 94 participants.
(88) The three different versions of the victim impact statements are reproduced in the Appendix. The only difference between the vignettes is the type of information presented about the emotional impact of the crime.
(89) In other words, we employed a between-subjects design in which the sole independent variable had three levels: Severe Emotional Injury, Mild Emotional Injury, and Control.
(90) Although the scale was marked with three-year increments, participants were instructed to indicate their opinion by marking an "X" anywhere on the scale. Thus, participants could recommend sentences that fell in between the three-year increments.
(91) Throughout this Article, "significantly" refers to statistical significance, which denotes the rejection of the null hypothesis--the possibility of no differences between the various groups--at a probability of less than 5%. Thus, "p" is defined as the probability of finding a difference or relationship between two groups as large as that observed if there were, in fact, no difference or relationship between them.
(92) Severe mean = 5.83; Mild mean = 3.43. t (132) = 11.47; p < .001.
(93) Severe mean = 6.16; Mild mean = 3.00. t (61) = 12.40; p < .001.
(94) Burglary: Control mean = 4.97. An analysis of variance (ANOVA) revealed significant overall differences: F (2, 203) = 53.62; p < .001. Robbery: Control mean = 5.20; overall F (2, 90) = 61.02; p < .001.
(95) In experimental psychology, this analysis is called a "manipulation check." This check ensures that the participants correctly perceive the variations in the severity of the harms. If participants did not correctly perceive these variations, any differences in the resulting sentences could not be plausibly related to differences in severity of injuries.
(96) t (121) = 2.81; p < .01; the Control mean did not differ significantly from the Severe mean (Control = 4.2 years). An ANOVA revealed a significant difference across all three conditions. F (2, 191) = 3.66; p < .05.
(97) t (60) = 2.49; p < .05; the Control mean did not differ significantly from the Severe mean (Control = 4.3 years). An ANOVA revealed a marginally significant difference across all three conditions. F (2, 89) = 3.07; p = .052.
(98) t (89) = 2.84; p < .01; the Control mean did not differ significantly from the Severe mean (Control = 4.94). An ANOVA revealed a marginally significant difference across all three conditions. F (2, 135) = 4.14; p = .05.
(99) F (2, 90) = 2.42; p = .095. Means: Severe = 5.28; Mild = 4.77; Control = 4.53.
(100) All Fs < 1.
(101) All Fs = 1.
(102) [chi square] (1) = 4.79; p < .05. Thus, although the severity of the victim's emotional injury did not have an observable overall effect on judgments about the seriousness of the crime, emotional injury did appear to influence some participants who were inclined to think that the crime was extremely serious.
(103) The sample consisted of laypersons, not judges. It is an open question whether judges imposing a sentence in a noncapital case would be similarly affected by different levels of emotional harm expressed in victim impact statements. However, evidence exists that judges, despite their training and expertise, rely on the same decision-making processes as laypersons, making them vulnerable to systematic mental shortcuts. See, e.g., Chris Guthrie et al., Inside the Judicial Mind, 86 CORNELL L. REV. 777 (2001). The sampling of laypersons (as opposed to judges) in this experiment is directly relevant in states in which juries are involved in sentencing. There are several states in which juries have the power to sentence in noncapital cases, although in some of these states, the jury's power is limited. See ARK. CODE ANN. [subsection] 5-4-103, 16-90-107 (Michie 1007 8c Supp. 2001); KY. REV. STAT. ANN. [section] 532.055(2) (Banks-Baldwin 2001); Mo. REV. STAT. [section] 557.036 (2000); OKLA. STAT. ANN. tit. 22, [subsection] 926-028 (West 1086 & Supp. 2002); TENN. CODE ANN. [subsection] 40-20-104, 40-20-107 (1997); TEX. CRIM. PROC. CODE ANN. art. 37.07 (Vernon 1981 & Supp. 2002); VA. CODE ANN. [section] 19.2-205.1 (Michie 2000 & Supp. 2002).
(104) See Guthrie et al., supra note 103.
(105) See Greene, supra note 63, at 333 (noting that some proponents of victim impact evidence argue that harm to victims "is directly relevant to gauging the severity of the crime and the appropriate punishment").
(106) See supra notes 5, 51 and accompanying text.
(107) See supra notes 21-27 and accompanying text.
(108) The example is similar to the one given by Justice White in his dissenting opinion in Booth v. Maryland, 482 U.S. 496, 516 (1987) (White, J., dissenting). The fact that a person may be subject to praise or blame based on an outcome caused by fortuitous circumstances is termed "moral luck," and has been discussed at length by moral philosophers. See, e.g., THOMAS NAGEL, MORTAL QUESTIONS 24-38 (1979).
(109) See supra notes 3-a and accompanying text.
(110) A number of criminal law theorists have argued that our intuitions are simply mistaken, and that punishment should not depend on moral luck. See, e.g., Joel Feinberg, Equal Punishment for Failed Attempts: Some Bad but Instructive Arguments Against It, 37 ARIZ. L. REV. 117, 118-19 (1995) (proposing the elimination of "the causal condition in the definition of all so-called completed crimes"); Sanford H. Kadish, Foreword: The Criminal Law and the Luck of the Draw, 84 J. CRIM. L. & CRIMINOLOGY 679, 679-81 (1994) (arguing that the "harm doctrine" is not rationally supportable because it "does not serve the crime preventive purposes of the criminal law, and is not redeemed by any defensible normative principle"). Thus, a person who attempts but fails to complete a crime deserves the very same punishment as the person who successfully completes the same crime. In contrast, Michael Moore and others adopt a version of retributivism that takes harm into account (even harm that is fortuitous) in gauging culpability and punishment. See MICHAEL MOORE, PLACING BLAME: A GENERAL THEORY OF THE CRIMINAL LAW 191-247 (1997). Moore calls this a "moral luck" form of retributivism. Michael Moore, Victims and Retribution: A Reply to Professor Fletcher, 3 BUFF. CRIM. L. REV. 65, 66 (1999).
(111) See Gail L. Heriot, The Practical Role of Harm in the Criminal Law and the Law of Tort, 5 J. CONTEMP. LEGAL ISSUES 145, 145-47 (1994); Stephen J. Morse, The Moral Metaphysics of Causation and Results, 88 CAL. L. REV. 879, 887-88 (2000).
(112) See Richard E. Nisbett & Timothy DeCamp Wilson, Telling More than We Can Know: Verbal Reports on Mental Processes, 84 PSYCHOL. REV. 231, 231-32 (1977).
(113) See Heriot, supra note 111, at 147; Moore, supra note 110, at 85-88.
(114) Elizabeth F. Loftus & John C. Palmer, Reconstruction of Automobile Destruction: An Example of the Interaction Between Language and Memory, 13 J. VERBAL LEARNING & VERBAL BEHAV. 585, 588 (1974).
(115) See id.
(116) Greene et al., supra note 13, at 154; see also Mark D. Alicke & Teresa L. Davis, The Role of a Posteriori Victim Information in Judgments of Blame and Sanction, 25 J. EXPERIMENTAL SOC. PSYCHOL. 362, 374 (1989) (finding that mock jurors punished the defendant more severely when the victim was characterized as "innocent" than when the victim turned out to be a dangerous criminal, even though the defendant had no information about the victim at the time of the incident).
(117) See Larry Alexander, Crime and Culpability, 5 J. CONTEMP. LEGAL ISSUES 1, 1-3 (1994).
(118) See Ugwuegbu & Hendrick, supra note 31, at 76.
(119) See id. at 78-79.
(120) See id. at 79.
(121) See id. at 84.
(122) See Ed Hayward, Lessons of Rink-Rage Case Touch All in Youth Sports, BOSTON HERALD, Jan. 26, 2002, at 5.
(123) See id.
(124) There is, of course, the familiar example of the victim who has an "eggshell skull." See George P. Fletcher, The Place of Victims in the Theory of Retribution, 3 BUFF. CRIM. L. REV. 52, 56 (1999) (noting that under the "eggshell skull" doctrine, "a defendant is liable for the unexpected and even unforeseeable consequences to the particular plaintiff'). But the frequency in the population of such victims is likely to be exceedingly low, making the physical harm heuristic reliable most of the time.
(125) The Supreme Court raised this concern in Booth v. Maryland, 482 U.S. 496, 505-06 (1987), and Justice Marshall also discussed the issue in his dissent in Payne v. Tennessee, 501 U.S. 808, 846 (1991) (Marshall, J., dissenting).
(126) See, e.g., Jonathan Baron & John C. Hershey, Outcome Bias in Decision Evaluation, 54 J. PERSONALITY & SOC. PSYCHOL. 569, 569-70 (1988).
(127) See id. at 571-72.
(128) Id. at 577.
(130) See supra note 53 and accompanying text (describing Justice Stevens's dissent in Payne); supra note 8 and accompanying text (discussing scholarly criticism of the use of victim impact statements).
(131) See, e.g., Richard S. Lazarus, The Cognition-Emotion Debate: A Bit of History, in HANDBOOK OF COGNITION AND EMOTION 3, 3 (Tim Dalgleish & Mick J. Power eds., 1999).
(132) See Graham et al., supra note 21, at 341 (discussing the results of an empirical study).
(133) In emotions literature, this desire is referred to as an "action tendency." RICHARD S. LAZARUS, EMOTION AND ADAPTATION 226 (1991).
(134) See Dacher Keltner et al., Beyond Simple Pessimism: Effects of Sadness and Anger on Social Perception, 64 J. PERSONALITY & SOC. PSYCHOL. 740, 750-51 (1993).
(135) See Craig A. Smith & Phoebe C. Ellsworth, Patterns of Cognitive Appraisal in Emotion, 48 J. PERSONALITY & SOC. PSYCHOL. 813, 833 (1985).
(136) See Keltner et al., supra note 134, at 751.
(137) See Jennifer S. Lerner et al., Sober Second Thought: The Effects of Accountability, Anger, and Authoritarianism on Attributions of Responsibility, 24 PERSONALITY & SOC. PSYCHOL. BULL. 563, 570 (1998).
(138) See Mark D. Alicke, Culpable Control and the Psychology of Blame, 126 PSYCHOL. BULL. 556, 568 (2000).
(139) See id.
(140) See id. at 566.
(141) This is a well known phenomenon in cognitive psychology in which decision makers search for confirming information rather than disconfirming information. See, e.g., P.C. WASON & P.N. JOHNSON-LAIRD, THE PSYCHOLOGY OF REASONING: STRUCTURE AND CONTENT 202-17 (1972). This phenomenon :has come to be known as "confirmation bias." Joshua Klayman & Young-Won Ha, Confirmation, Disconfirmation, and Information in Hypothesis Testing, 94 PSYCHOL REV. 211, 211 (1987). The blame-validation hypothesis holds that a negative affective reaction to a situation--termed a "spontaneous evaluation"--leads people to engage in a biased search for information. See Alicke, supra note 138, at 564-66.
(142) Empirical evidence supports the notion that jurors seek evidence to confirm judgments they may have made earlier in the case. In one study, mock jurors read evidence from a criminal case, loosely based on the Bernard Goetz case, in which researchers portrayed four young men who were victims of a shooting on a subway as either gang members or star athletes. See Alicke, supra note 138, at 567. Mock jurors were told that they could hear from only three of four witnesses before making their decision about the defendant's guilt, and they were asked to choose which witnesses they would hear from. Id. Two of the witnesses were pro-prosecution and the other two were pro-defense. Id. Mock jurors who learned that the victims were star athletes chose to hear from the prosecution witnesses more often than mock jurors who were told that the victims were gang members. Id. Mock jurors, having already decided how severely to punish the defendant early in the case, sought out the kind of evidence that would confirm the judgment they had previously made. See id.
(143) See supra note 52 and accompanying text.
(144) See Paul Rozin et al., The CAD Triad Hypothesis: A Mapping Between Three Moral Emotions (Contempt, Anger, Disgust) and Three Moral Codes (Community, Autonomy, Divinity), 76 J. PERSONALITY & SOC. PSYCHOL. 574, 574-76, 585 (1999).
(145) See supra note 135 and accompanying text.
(146) Rozin et al., supra note 144, at 576.
(147) See id.
(148) See id. at 575-76.
(149) See id.
(150) See Logan, supra note 8, at 161-62; Katie Long, Note, Community Input at Sentencing: Victim's Right or Victim's Revenge?, 75 B.U. L. REV. 187 (1995).
(151) See Logan, supra note 8, at 153-56.
(152) Jonathan Haidt, The Moral Emotions, in HANDBOOK OF AFFECTIVE SCIENCES (R.J. Davidson et al. eds., forthcoming 2002), available at http://wsrv.clas.virginia.edu/~jdh6n/ themoralemotions.html.
(153) For a similar argument, see Luginbuhl & Burkhead, supra note 66, at 5-6 (noting that jurors may perceive imposing severe punishment as an indirect way of helping the victim's family).
(154) See Payne v. Tennessee, 501 U.S. 808, 830 n.2 (1991) (noting that Payne did not affect the Court's holding in Booth that the admission of a victim's family member's opinion regarding the appropriate sentence violated the Eighth Amendment).
(155) See supra notes 96-99 and accompanying text.
(156) Booth v. Maryland, 482 U.S. 496, 506 n.8 (1987), overruled by Payne, 501 U.S. at 827.
(157) Payne, 501 U.S. at 823.
(158) See id. at 825-26.
(159) Id. at 825.
(160) See, e.g., REID HASTIE & ROBYN M. DAWES, RATIONAL CHOICE IN AN UNCERTAIN WORLD: THE PSYCHOLOGY OF JUDGMENT AND DECISION MAKING 56 (2001) (noting that the addition of experienced clinical judgment led to less reliable diagnoses about psychosis compared to the use of personality test scores alone); Robyn M. Dawes, A Case Study of Graduate Admissions: Application of Three Principles of Human Decision Making, 26 AM. PSYCHOLOGIST 180, 182, 186-87 (1971) (indicating that the addition of personal interview information led to less reliable graduate school admission decisions compared to the use of grades and test scores alone).
(161) One example of a specific cognitive heuristic that leads to systematic over or underweighting of information is the availability heuristic, which suggests that we base the probability of a future event on relevant examples that are cognitively accessible. That is, we overweight examples that come to mind first. For example, people who live on flood plains are less likely to buy flood insurance if floods have not occurred in the recent past. See PAUL SLOVIC, THE PERCEPTION OF RISK 39-40 (2000).
(162) See Dawes, supra note 160, at 187.
(163) See HASTIE & DAWES, Supra note 160, at 64-65.
(164) See ROBYN M. DAWES, HOUSE OF CARDS: PSYCHOLOGY AND PSYCHOTHERAPY BUILT ON MYTH 86-89 (1994).
(165) See id.
(166) See id.
(167) See John S. Carroll et al., Evaluation, Diagnosis, and Prediction in Parole Decision Making, 17 LAW & SOC'Y REV. 199, 215-17 (1982).
(168) Id. at 222.
(169) Id. at 219.
(170) See supra Part IV.A-B.
(171) See supra Part IV.A.
(172) See supra Part IV.B.
(173) See supra notes 134-46 and accompanying text.
(174) See supra notes 124-25 and accompanying text.
(175) See Bandes, supra note 8, at 392-93.
(176) See Harris, supra note 8, at 91-92.
(177) See Bandes, supra note 8, at 390-91; Harris, supra note 8, at 91.
(178) See Andy Pasztor, TWA Crash Probe Leaves Puzzles, Stricter Standards, CHI. TRIB., Mar. 4, 2001, [section] 12, at 1.
(179) Aircraft Accident Report: In-flight Breakup over the Atlantic Ocean, Trans World Airlines Flight 800, National Transportation Safety Board, at http://www.ntsb.gov/Publictn/2000/ aar0003.htm (last visited Nov. 17, 2002).
(180) See Greene, supra note 63, at 332.
(181) See Payne v. Tennessee, 501 U.S. 808, 830 n.2 (1991).
(182) For an excellent comprehensive review of these practices nationwide, see Logan, supra note 8, at 151-56.
(183) For example, New Jersey permits only one family member to provide victim impact testimony in a capital trial. See State v. Muhammad, 678 A. 2d 164, 180 (N.J. 1996). This rule limits the duration of the jury's exposure to victim impact testimony and therefore might limit the powerful influence of the attributional and emotional decision biases discussed earlier. Maryland bars live victim impact testimony, and instead requires that victim impact information be in written form, incorporated in the presentence investigation report. MD. ANN. CODE art 41, [section] 4-609(d)(1997).
Janice Nadler, Research Fellow, American Bar Foundation; Assistant Professor of Law, Northwestern University School of Law.
Mary P. Rose, Research Fellow, American Bar Foundation (while this research was conducted); Assistant Professor of Sociology and of Law, University of Texas at Austin.
|Printer friendly Cite/link Email Feedback|
|Author:||Nadler, Janice; Rose, Mary R.|
|Publication:||Cornell Law Review|
|Date:||Jan 1, 2003|
|Previous Article:||The capital jury and empathy: the problem of worthy and unworthy victims.|
|Next Article:||The cognitive components of punishment.|