Social Scientific Analyses of Contemporary Racism

The various theoretical approaches to understanding racial bias can be arranged along a continuum that ranges from an emphasis on broad, social structural influences, to a focus on dynamic social and group-based processes, to more individualistic and intrapsychic explanations. At one end of this continuum, sociological theorists underscore the broader structural settings in which racism occurs and argue that these macro variables are virtually always implicated in the creation and perpetuation of racialized individual behaviors. Thus, Reference Bonilla-SilvaBonilla-Silva (1997) argues that understanding the instances, attitudes, or ideologies that constitute “racism” requires an examination of the many social systems that are based in part on racial hierarchies, one that acknowledges the various ways that these hierarchies shape social relations among groups that are situated differently within them.

In the case of race relations in the United States, in particular, the historical pattern of white domination has ensured inequitable distributions of resources and opportunities. Yet the methods by which these inequalities are preserved have changed shape over time, making them difficult to identify and even harder to rectify. Thus, racial hierarchies and inequalities, which give rise to prejudice and discrimination, are operated in seemingly even-handed, “neutral” ways (e.g., Reference Haney and HurtadoHaney & Hurtado 1994) and are produced and reproduced at a systemic level that makes them appear “natural” and the explicit “fault” of no particular individual (see also Reference MasseyMassey 2007; Reference Haney LópezHaney López 2000; Reference WacquantWacquant 2000).

At the group level, a number of social psychologists have examined forms of intergroup discrimination, focusing on the role of social identity, as a way to understand contemporary racism. This research has acknowledged the dynamic nature of prejudice and identifies individuals' salient group membership(s), which can be multiple, contradictory, and fluid over time and across situations, as the source of in-group favoritism (e.g., Reference ScheepersScheepers et al. 2006). However, because some group identities are more stable than others, certain demographic characteristics (such as gender and racial or ethnic identity), which tend to be less malleable, are more centrally related to prejudiced beliefs. This line of research suggests that in-group bias is not necessarily manifested through the amplification of negative attributions directed at out-groups, but through the assignment of more positive attributes for in-group members (Reference Dovidio, Blanchard and CrosbyDovidio et al. 1989), explicit expressions of in-group favoritism (Reference TajfelTajfel 1982), and the withholding of empathy or positive affect for out-group members (Reference Pettigrew and MeertensPettigrew & Meertens 1995).

In addition, social group stereotyping in particular is understood to be shaped by sociohistorical factors, which accounts for the persistence of widely known (if not widely accepted) culturally specific stereotypes about different groups in any given society (Reference DevineDevine 1989; Reference Devine and ElliotDevine & Elliot 1995). As a result, racial and cultural stereotypes are generally highly accessible to most members of a given culture and can shape views and perceptions about different groups in a subtle yet pervasive manner.

At a more individual level, social psychologists have emphasized the role of certain cognitive processes—including the human tendency to rely on stereotypes and other heuristics to simplify a complex social world—that may contribute to contemporary racism. Cognitive shortcuts, such as stereotyping, allow for more efficient information processing, memory, and retrieval of the complicated stimuli with which we are bombarded on a day-to-day basis (Reference Fiske and TaylorFiske & Taylor 1991). Unfortunately, of course, these very shortcuts often produce overgeneralized, inaccurate knowledge about the social world (Reference KriegerKrieger 1995). Some scholarship on stereotyping distinguishes prejudice, which is characterized as more affective and motivational, from the more cognitive and somewhat automatic and nonconscious processes that are sometimes at the basis of the stereotyping process (Reference Amodio and DevineAmodio & Devine 2006). Nonetheless, both prejudice and stereotyping can contribute to biased assessments and, consequently, lead to discriminatory behavior.

Reference Gaertner, Dovidio, Dovidio and GaertnerGaertner and Dovidio's (1986) theory of aversive racism integrates some of these perspectives, joining the cognitive and affective aspects of stereotyping with an appreciation of the situational and structural factors that help activate forms of racial discrimination. They argue that although blatant, intentional racism has become less socially acceptable in the post–civil rights era, a more subtle, unconscious form of prejudice has become increasingly prevalent. Indeed, they suggest that, unlike like other forms of racism that have been associated with political conservatism, aversive racism is particularly prevalent among “well intentioned, liberal, and well educated individuals” who espouse egalitarian beliefs (Reference Hodson and ChinHodson et al. 2004:120). Aversive racists are more likely to engage in forms of prejudiced or discriminatory behavior when their behavior can be justified on nonracial grounds (Reference DovidioDovidio 2001), and they often express their prejudice by showing bias in favor of their own in-group rather than outward bias against a group of disfavored others (Reference Hodson and ChinHodson et al. 2004).

In addition to its intrapsychic dimensions, aversive racism is thought to be triggered by certain aspects of the context or situation in which it occurs. Thus, this particular form of racism is more likely to be activated in settings where the prevailing social norms are conflicting or ambiguous, or in situations where the persons making judgments do not have adequate time or sufficient cognitive resources to fully process the available information (Reference DevineDevine 1989; Reference KawakamiKawakami et al. 1998). Conversely, the activation of aversive racism is less likely in situations where race itself is made salient, or where there are no race-neutral reasons available to explain disparate treatment (Reference Gaertner and DovidioGaertner & Dovidio 2005; Reference SommersSommers 2007).

Racism and Death Sentencing

There are a number of reasons why racial animus might be expected to operate with greater force in the criminal justice system—where the explicit mandate is to administer punishment to wrongdoers—than in the society at large. For one, it is easier to punish persons more harshly if they have already been demonized, are perceived as somehow less than fully human, or are regarded as fundamentally “other” (e.g., Reference Bandura and ReichBandura 1989). The derogated status of criminals in general lowers the social inhibition against subjecting them to harsh punishment, helping in part to justify—psychologically as well as legally—the treatment they receive. At the same time, it may be easier to believe that people from already disfavored groups—here, those who are racially stigmatized—have done bad things and, therefore, deserve punishment. This is especially true if, as is the case with racial stereotypes in our society, belief in their inherent “badness” is part of their stigma.

In addition, belief in the inherent, pre-existing badness of certain groups may make it easier for dominant group members to ignore flaws in the legal processes by which those derogated others are officially blamed and punished for their transgressions. Similarly—because they are more feared and despised to begin with—it may be easier to exaggerate the seriousness of the things that members of already disfavored groups have been found guilty of doing. That is, it may be easier to believe that whatever crimes these “others” have committed are per se more heinous (whereas if they were committed by members of one's own group, they would be less so). Finally, because it is more difficult to identify with persons perceived as “other,” dominant group members can more easily distance themselves from the pains of whatever punishment these “others” receive, notwithstanding the possibility that such punishment may be unjustly administered or excessive in amount.

Within this larger criminal justice system context, death penalty trials are in many ways ideal settings for contemporary racism to flourish. For one, because the death qualification process systematically excludes persons on the basis of their strong death penalty views from the jury, the demographic makeup of the capital jury is distinctive and problematic.Footnote ¹ That is, compared to juries seated in nondeath cases, death-qualified jury pools are disproportionately white, male, older, and more religiously and politically conservative (Reference Filkins and TindaleFilkins et al. 1998; Reference GrossGross 1998; Reference HaneyHaney et al. 1994; Reference Lynch and HaneyLynch & Haney 2000). These demographic characteristics tend to correlate with measures of subtle racism (Reference BarnesBarnes 1997; Reference EmersonEmerson et al. 1999; Reference Meertens and PettigrewMeertens & Pettigrew 1997). Furthermore, survey research suggests that support for the death penalty among whites is highly correlated with measures of anti-black racial prejudice and stereotyping (Reference ArthurArthur 1998; Reference Barkan and CohnBarkan & Cohn 1994; Reference Bobo and JohnsonBobo & Johnson 2004; Reference SossSoss et al. 2003; Reference Unnever and CullenUnnever & Cullen 2007).

Second, the kinds of crimes that give rise to capital prosecutions not only provoke heightened levels of anger and fear among jurors but also are more likely to activate cultural stereotypes about minority ethnic and racial groups. In this regard, Reference Quillian and PagerQuillian and Pager (2001) have argued that the stereotype of blacks as violent and criminally inclined is one of the most pervasive, well-known, and persistent stereotypes in American culture. While other negative cultural stereotypes about blacks have significantly diminished, this one has remained strong and influential, particularly among whites (see also Reference Devine and ElliotDevine & Elliot 1995; Reference Peffley and HurwitzPeffley & Hurwitz 2002; Reference Hurwitz and PeffleyHurwitz & Peffley 1997).

Moreover, research suggests that persons tend to judge those defendants charged with stereotype-consistent offenses more harshly than those charged with stereotype-inconsistent crimes (Reference GordonGordon 1990; Reference Jones and KaplanJones & Kaplan 2003; Reference Mazzella and FeingoldMazzella & Feingold 1994). In the specific context of capital cases, Reference Fleury-SteinerFleury-Steiner (2004) found that stereotypes among capital jurors about minority defendants and their propensity toward violence shaped deliberations and outcomes. Research also indicates that there is variable discrimination against black capital defendants who kill white victims as a function of the stereotypicality of their appearance (Reference EberhardtEberhardt et al. 2006). Using defendant photographs from actual capital cases tried in Philadelphia, Eberhardt and colleagues found that the more “stereotypically black” the capital defendant appeared to independent raters, the more likely it was that the defendant had received a death sentence. Even among criminal justice professionals, racial stereotypes, when activated, have been shown to have an adverse effect on assessments of offenders' culpability and the severity of punishment deserved (Reference Graham and LoweryGraham & Lowery 2004).

Additional aspects of the context in which capital cases are decided render them especially susceptible to the influence of race. Capital jurors are placed in a unique and unfamiliar situation where the decisionmaking norms and the values that should be brought to bear on the task at hand are particularly unclear. Key elements of the death penalty decision—including making judgments about moral culpability and blameworthiness, gauging the heinousness of the capital crime, and weighing past acts of violence against the defendant's worth or value as a person—are inherently subjective assessments whose parameters are both unfamiliar and emotionally wrought. Research suggests that racial animus may be more readily mobilized and acted upon in a setting where decisionmaking norms are so ambiguous.

In addition, capital jury sentencing instructions are notoriously difficult for jurors to understand and apply (Reference Eisenberg and WellsEisenberg & Wells 1993; Reference Haney and LynchHaney & Lynch 1994, Reference Haney and Lynch1997; Reference Luginbuhl and BurkheadLuginbuhl & Burkhead 1994; Reference TiersmaTiersma 1995; Reference WienerWiener et al. 1995), which increases the likelihood that their judgments will be shaped by pre-existing biases. Indeed, prior research indicates that when persons rely on rules or guidelines that are unclear, or difficult to understand and follow, their judgments are more likely to be influenced by race (Reference BodenhausenBodenhausen 1988; Reference Bodenhausen and LichtensteinBodenhausen & Lichtenstein 1987; Reference Bodenhausen and WyerBodenhausen & Wyer 1985; Reference Hill and PfeiferHill & Pfeifer 1992; Reference Pfeifer and OgloffPfeifer & Ogloff 1991; Reference RectorRector et al. 1993). In fact, researchers have found that poor juror comprehension of capital sentencing instructions is related to racially biased death sentencing (Reference Lynch and HaneyLynch & Haney, 2000; Reference Shaked-SchroerShaked-Schroer et al. 2008).

Moreover, because the creation of empathy for the defendant—usually through the presentation of a mitigating narrative in the penalty phase of the case—is important in convincing jurors to vote for life over death (Reference HaneyHaney 2004), black defendants may be especially disadvantaged at the hands of typically white jurors. We know that empathy is more likely to be felt for those who are perceived as similar to one's self (Reference LinderLinder 1996), and for this reason it is less likely to be extended to defendants who are perceived as fundamentally “other,” not only because of what they have been convicted of doing but also because of their racial characteristics (Reference EisenbergEisenberg et al. 1996; Reference HaneyHaney 2004, Reference Haney2005).

Finally, and in a related way, the racial divisions that plague American society operate to subject black capital defendants to what has been called “biographical racism”—the “accumulation of race-based obstacles, indignities, and criminogenic influences that characterizes the life histories of so many African American capital defendants” (Reference HaneyHaney 2004:1557). The fact that they are more often exposed to criminogenic forces and factors over the course of their lives helps account for their presence as defendants in capital cases. Yet most white jurors have no framework for understanding the operation of these forces. Thus, black capital defendants may be doubly disadvantaged—first by being subjected to social structural forces that shape their lives in problematic and even destructive ways, and second by having the nature and value of those lives judged finally by persons ill equipped to understand their structural determinants and mitigating aspects.

Thus, all of these forces and factors—demographic and attitudinal features of capital jurors, the characteristics of capital cases, the parameters of the penalty-phase decisionmaking process, and the larger social and institutional structure that facilitates systemic racism—likely play important roles in the activation and expression of racial bias in capital jury decisionmaking.

Evidence of Persistent Race-Based Death Sentencing

In light of the many reasons to expect that racial factors would operate with particular force and effect in the administration of capital punishment, it is not surprising to find a breadth of evidence documenting the extent to which they have and do. Indeed, in the long history of the death penalty in the United States, blacks have been consistently sentenced to die in numbers that greatly exceed their percentages in the population at large. Even before the American Revolution, the Southern colonies managed their captive slave workforces in part by relying on “ever-increasing lists of capital statutes” that were applied to blacks and not whites (Reference BannerBanner 2002:8). The legacy of this early system persisted to modern times. For example, when the U.S. Supreme Court examined the operation of the nation's system of death sentencing in Furman v. Georgia (1972), Justice Thurgood Marshall noted that a total of 405 blacks had been executed for the crime of rape over the preceding 40 years, compared to 48 whites (1972:365, n. 147).

Although the problem of racialized death sentencing was discussed at some length in Furman, it was the standardless discretion exercised by capital juries that the Court found decisive in declaring state death penalty statutes unconstitutional as applied. Following the Court's reinstatement of capital punishment four years later (Gregg v. Georgia 1976), evidence that race still played an important role in the imposition of the death penalty continued to mount. Studies indicated that blacks were overpunished as defendants or undervalued as victims, or both, in capital cases (e.g., Reference Bowers and PierceBowers & Pierce 1980; Reference Gross and MauroGross & Mauro 1984, Reference Gross and Mauro1989; Reference ZieselZiesel 1981). By the time the Court considered the issue of racially discriminatory death sentencing directly, in McCleskey v. Kemp (1987), careful data analyses that controlled for a host of potentially confounding variables (e.g., Reference Baldus, Woodworth and PulaskiBaldus, Woodworth, & Pulaski 1990) revealed many of the same disparities.

Although the McCleskey Court ruled that statistical patterns of racial disparities in death sentencing alone lacked constitutional significance, those patterns were repeatedly documented in the social science literature. In a congressionally mandated summary and review of some 28 studies of the topic, a Government Accounting Office report found that the race of the victim significantly influenced death sentencing in about four out of five of the studies, and that the race of the defendant significantly influenced it in about half of them (U.S. Government Accounting Office 1990). Thus, despite some variation in the outcomes of the studies, racial factors continued to play a significant role in the death sentencing process, sometimes strikingly so, as when a black capital defendant had been convicted of a potentially capital murder in which the victim was white.

More recently, studies of capital jury sentencing verdicts rendered during the 1980s and 1990s in the state of New Jersey and in the city of Philadelphia found that juries were significantly more likely to sentence black defendants to death in comparison to all other defendants (Reference Baldus, Woodworth and ZuckermanBaldus, Woodworth, Zuckerman, et al. 1998).Footnote ² The Philadelphia study examined all potentially capital cases from 1983 to 1993 and indicated that a black capital defendant was 9.3 times more likely to be sentenced to death by the jury than a similarly situated defendant of another race or ethnicity. While these studies demonstrated a race of victim effect as well—cases involving black victims produced fewer death sentences than others—this effect was weaker than the influence of race of defendant. Similarly, an examination of capital case outcomes in the state of Maryland found that, compared to all other combinations of victims and offenders, blacks who killed white victims were significantly more likely to have death notices filed against them, proceed to trial as capital cases, and receive death sentences at the penalty phase of their case in comparison to otherwise similarly situated defendants (Reference Paternoster and BramePaternoster & Brame 2003).

The results of these archival studies parallel findings obtained from experimental or “simulation” research on race and sentencing. This approach permits greater control over extraneous variables, so that only the race of the defendant and/or the race of the victim can be systematically varied. In one early meta-analysis of existing experimental studies on race-of-defendant effects, Reference Sweeney and HaneySweeney and Haney reported that, despite some inconsistent findings in individual studies, there was a significant overall race-of-defendant effect such that black defendants tended to be sentenced more harshly than others. In addition, they found that “studies demonstrating greater methodological rigor more consistently uncovered racial bias in sentencing decisions” (1992:191).

In later research using an experimental or simulation design, the present authors found significant race-of-defendant effects in individual-level death penalty sentencing verdicts (Reference Lynch and HaneyLynch & Haney 2000). In that study, a total of 348 jury-eligible adult participants individually viewed one of four identical versions of a staged capital penalty trial that varied only by the race of the defendant and the victim (black or white). Participants then determined the sentence—either life without possibility of parole or death. Our findings indicated that the black defendant was significantly more likely to be sentenced to death than the white defendant and that the discrepancy in sentence outcome was particularly pronounced in the cross-racial conditions.Footnote ³

Even more recently, we replicated this initial study in another jurisdiction and modified the design to include a group deliberation component (Reference Lynch and HaneyLynch & Haney 2009). More than 500 jury-eligible, death-qualified adults were randomly assigned to four- to seven-person “juries” that viewed one of the four versions of the penalty trial tape and deliberated to determine the sentence. We found that the black defendant was again more likely to be sentenced to death than the white defendant, and this effect was amplified by the race of the victim (such that, for example, the white defendant convicted of killing a black victim received the fewest death verdicts). Notably, we also found that the race-of-defendant effect was significantly strengthened by the deliberation process, which tended to increase our juror-participants' support for a death sentence overall and especially so in the black defendant conditions.

Capital Juror Characteristics and Racialized Death Sentencing

Obviously, these broad patterns of racially discriminatory death sentencing result from the aggregate thoughts and actions of individual decision makers—capital jurors—who make judgments and render verdicts in tandem with others. Reference Sweeney and HaneySweeney and Haney concluded their meta-analysis of experimental research on racially discriminatory sentencing by emphasizing precisely this interplay between defendant, victim, and juror characteristics: “Thus sentencing discrimination is not generic bias against Black defendants. Rather, it appears to result from a specific punitive reaction to crimes committed by members of another race against persons of one's own” (1992:192). In recent years, increased attention has been focused on the specific characteristics of capital jurors that may help explain the racialized nature of their specific punitive reactions.

The work of the Capital Jury Project (CJP) on this topic is especially noteworthy. CJP is a multistate study in which a large sample of actual capital jurors who had previously served through the penalty phase of a death penalty trial are systematically interviewed about their experiences as jurors. In one CJP study that analyzed juror data from 14 different states, Reference BowersBowers et al. (2001) found what they termed a “white male dominance” effect in cases where black defendants were accused of killing white victims. Specifically, when the jury included five or more white men, the jury was significantly more likely to sentence the defendant to death than when it included four or fewer white men (71 percent versus 30 percent ending in death sentences). Conversely, the same study also uncovered what the researchers termed a “black male presence” effect, such that the presence of one or more black men on the jury significantly reduced the likelihood of a death sentence in the case. In another CJP study utilizing data from capital cases tried in South Carolina, Reference GarveyGarvey found that black jurors were much more likely to “keep the sin separate from the sinner” (2000:47) in both black and white defendant cases and had more overall positive feelings toward the defendant than did white jurors.

With the CJP insights about the importance of considering the specific racial makeup of the capital jury in mind, we returned to our latest study of discriminatory death sentencing (Reference Lynch and HaneyLynch & Haney 2009), and reanalyzed a significant subset of our data to determine the role of these jury-based interpersonal dynamics in producing our results. We also sought to extend understandings of the psychological processes by which the white male dominance effect comes about. That is, how and why do white male jurors reach the racially tinged sentencing verdicts that they do, and what are the patterns by which their private judgments influence the views of their fellow jurors?

The Present Research

In the remainder of this article, we specifically focus on our findings regarding the interactions between juror characteristics and case factors, as mediated by the deliberation process.Footnote ⁴ As noted above, we found that white male jurors differed from all others (i.e., women and nonwhite men) in their sentencing decisions as a function of defendant race (Reference Lynch and HaneyLynch & Haney 2009). Our goal here is to explore in much more depth the ways in which the two demographic groups of jurors (white men and nonwhites/women) differed from each other, and within their own groups, in the way they processed the evidence that was presented in the penalty trial, and in their perceptions of the defendant, as a function of the racial dynamics that were present in the trial they viewed. Notwithstanding the obvious limitations of simulated trial studies in terms of the reduced gravity, emotionality, and complexity that can be created in an experimental setting, this method offers the best way to control for literally all extraneous factors that may shape juror decisionmaking and confound potential race effects. It also allows us to measure a number of aspects of the decisionmaking process that would otherwise be impossible to do in a real capital case, including the deliberation process, individual participants' perceptions, cognitions and emotions in relation to the case, and pre- and postdeliberation assessments.

We first describe the specific elements of our methodological design and then detail our findings in several subsections. We begin with descriptive data to provide an overview of our participants and their characteristics, including the demographic and attitudinal differences between our participant subgroups of interest. We follow with a delineation of jury group-level findings, including how the demographic composition of those small groups appeared to interact with racial characteristics of the case in shaping verdicts. We then provide a detailed examination of how white men differed from nonwhites and women in their individual sentencing decisions, as a function of the race of the defendant, followed by our sets of findings on how these two subgroups interpreted specific evidence in the case and how they viewed the defendant's attributional and motivational “profile.” Following our results sections, we interrogate whether, in light of these and other findings about the pervasiveness of racial bias in the administration of capital punishment, a fair and equitable death penalty system is achievable, and we consider how our findings provide additional insight into our understanding of contemporary racism.

Overview of the Method and Design

The study itself was designed to minimize some of the weaknesses that are often found in jury simulation studies. Specifically, we used jury-eligible nonstudent participants, relied on audio and visual stimulus materials (rather than paper transcripts or summaries), and, most important, included a group deliberation component which took place in a non-university setting that mimicked more closely a jury room. The experiment was conducted in a large, relatively urban county in California with nonstudent adults who were screened in advance for their jury eligibility and death qualification status.

A total of 539 participants were selected and assigned to one of 100 small group (four- to seven-person) juries. We aimed for six-person jury units by assigning seven people to any given jury slot. If fewer than four people showed up for a time slot, the session was canceled, the participants who came were paid, and they were then offered an opportunity to sign up for a different session. Each jury unit viewed a video of one of four versions of a capital penalty trial, which were identical in all aspects except that the race of the defendant (either black or white) and the race of the victim (black or white) were varied to create the four experimental conditions.

The simulated capital penalty trial included witness testimony, opening statements and final arguments from the defense and prosecution, and the judge's reading of the set of California pattern instructions to be used in sentence determination. The robbery-murder case facts were based on an actual California capital penalty trial, chosen because it contained a number of features of a “typical” capital murder and also because pre-testing had shown that it contained “mid-range” case facts—ones that were neither so aggravated nor so mitigated that participants would vote overwhelmingly for one or another sentencing verdict. The prosecution's penalty-phase aggravation consisted of the circumstances of the crime, victim impact testimony, evidence that the defendant had been accused of a prior similar crime, and the defendant's apparent lack of remorse for the present crime. Defense mitigation included evidence that the defendant had been severely abused as a child, had suffered from alcohol and drug abuse at the time of the crime, and had pre-existing psychiatric problems that may have impaired the defendant's mental state, and testimony that the defendant was still loved and needed by his family. We pre-tested the case to make sure it was perceived as neither so aggravated that it nearly always ended in death, nor so mitigated that it primarily ended in a life verdict.

After jurors finished viewing the videotape, but before they had an opportunity to discuss the case, they were asked to confidentially and individually make a written verdict choice and to indicate their confidence in that choice. Participants were told that this was a nonbinding “straw vote” that they were free to change during or after deliberation. Each group of jurors was asked to select a foreperson and then to deliberate to reach a unanimous sentencing verdict of either life without the possibility of parole or the death penalty. The jury was also asked to indicate its collective confidence in the decision on a 7-point scale. Each jury was given copies of the judicial instructions that had been read by the judge and a verdict form on which to record the sentencing decision.

A video camera recorded deliberations, and the jurors were left alone in the jury room to deliberate. The deliberation period was limited to a maximum of one and a half hours. If the jury was unable to reach a unanimous verdict within that time limit it was declared deadlocked, and the nature of the split was noted on a “mistrial voting form,” including each individual juror's verdict and confidence rating. When the deliberation was completed and the instructions and the verdict forms had been collected, each individual juror was asked to complete a series of questionnaires, including a “juror perceptions” questionnaire that contained several open-ended questions about the decisionmaking process as well as a series of closed-ended questions that asked for evaluations of the defendant, attorneys, witnesses, penalty trial evidence, and arguments presented, and perceptions about why the defendant committed the capital crime. In addition, participants completed an instructional comprehension questionnaire and a demographics questionnaire.

Results

Jury Unit Demographics, Race of Defendant, and Sentence Decision

This study was designed to examine individual-level effects, nested within jury groups, so it did not include the large number of jury groups that would be needed to achieve the required statistical power when using the jury as a single unit of analysis.Footnote ⁶ So although most of our analyses rely on individual-level measures (i.e., demographics, perceptions, case assessments, instructional comprehension, etc.), we report group-level data for their descriptive value. Thus, we note first that the majority of the jury groups favored a death sentence—65 of the 100 jury units voted unanimously or in the majority for death after deliberation. The remainder were either evenly split (6) or unanimously or majority/unanimously in favor of a life sentence (29). This preference for death was not necessarily predictable based on the straw vote ballots. Of the 100 groups, only 47 percent went into deliberations with a majority of the group favoring death. As prior research on jury group dynamics has indicated, majority at first ballot typically predicts final verdict (Reference HansHans 2007; Reference Sandys and DillehaySandys & Dillehay 1995). Moreover, there were not significant differences in the straw vote majority compositions as a function of defendant race; in fact, only 44 percent of the black defendant cases favored death at the straw ballot, whereas 50 percent of the white defendant cases had a majority favoring death at the same stage.

As Figure 1 illustrates, there were differences in final case outcomes as a function of defendant race and ratio of white men on the jury. We created a dichotomous variable that split the juries at the median, so those with 33 percent or more white males were considered “high” white male juries (this designated a proportion of white male representation that, at its lowest, was just above their overall representation among our participants). In practical terms for our four- to seven-person small groups, this meant that a minimum of two white men were necessary to qualify as having a high concentration. Our logic for making such a split is based on Reference BowersBowers et al.'s (2001) implication that there may be a nonmajority tipping point of white male concentration on the capital jury that triggers biased influence on the group.

Figure 1. Percentage of Majority or Unanimous Death Sentence Outcomes by Defendant Race and White Male Composition of Jury.

Looking only at the cases where there was a unanimous or majority verdict (N=94), we found that the juries with a high concentration of white men were much more death-prone toward black defendants than were juries with a low concentration of white men, or ones that viewed the white defendant. As is illustrated in Figure 1, fully 86 percent of those heavily white male juries in the black defendant condition favored death verdicts. Otherwise, the death percentage ranged from 60 to 68 percent. Independent samples t-tests comparing high and low white male group composition revealed that mean differences for group verdict in the black defendant condition were marginally significant, favoring a death verdict (t=−1.76, p<0.09), whereas in the white defendant condition, there were no differences between high and low white male jury groups. We also looked at whether the ratio of white men on the jury predicted verdict severity for the black defendant and the white defendant, and found that it had no impact for the white defendant (β=−0.123; t=−0.857; R ²=0.01). White male ratio did, however, predict sentence severity for the black defendant to a marginally significant degree (β=0.271; t=1.95, p=0.057; R ²=0.07).

White men were disproportionately likely to become forepersons, which may have resulted in an even more substantial influence during deliberation. Specifically, although white men made up 32 percent of the participants, they were forepersons on 45 of the 100 mock juries. This disproportion is more striking in light of the fact that 18 of the mock juries had no white men at all (in contrast, only one jury was fully constituted by white men).Footnote ⁷ Put somewhat differently, then, on the 81 juries in which white men had the opportunity but were not compelled to be the foreperson, they represented only 37 percent of participants but 54 percent of the forepersons. Of the 16 juries that had both a white male foreperson and a black defendant, only two (12.5 percent) favored life after deliberation. In contrast, 31 percent of the white defendant cases with white men as forepersons favored a life sentence after deliberation.

Race and Gender of Participant, Race of Defendant, and Sentence Decision

We conducted a number of individual-level analyses and found that (among other findings) both gender and race/ethnicity of our participants interacted with the defendant race independent variable in terms of sentence outcomes, and in terms of how case evidence was evaluated. As in our previous experiment, no main race-of-victim main effect on sentence outcome was obtained in this study across our participants or as an interaction effect with gender/race of juror.

Overall, at the individual level, approximately 55 percent of our participants indicated a preference for death in their straw votes,Footnote ⁸ and after deliberation, about 66 percent favored death, indicating that a punitive shift resulted from the deliberation process. Men were significantly more likely to vote for death in their straw votes than women (60 percent for men versus 51 percent for women; Chi-square (1)=3.94, p<0.05); that gender difference disappeared for final votes (68 versus 65 percent for death). White participants were also significantly more likely to cast straw votes and final verdicts for death than nonwhites, although the gap shrank after deliberation (58 versus 39 percent at the straw vote; Chi-square (1)=10.69; p=0.001, and 69 versus 54 percent at the final vote; Chi-square (1)=8.17, p<0.01). White men as a distinct subgroup were significantly more likely to vote for death at the straw vote than the others (64 versus 50 percent; Chi-square (1)=8.71, p<0.01); however, that gap narrowed to marginal significance by the final vote (72 versus 64 percent; Chi-square (1)=3.15; p<0.10). These data indirectly suggest that white men persuaded the other jurors to move closer to their own death-leaning position during the deliberation process.

We also examined whether and how our two distinct groups of jurors made decisions as a function of the race of the defendant. We found that white male participants assigned a significantly higher percentage of death sentences than the women and nonwhite participants to the black defendant at the straw vote (Chi-square=7.48, p<0.01) and in their final verdict (Chi-square=9.50, p<0.01), but not to the white defendant at either judgment point (see Figure 2). In terms of within-group race-of-defendant effects, we also found that white men were significantly more likely to select a death sentence as the final verdict for the black defendant than for the white defendant, whereas women and nonwhite participants (both as distinct groups and as a single combined group) did not demonstrate significant differences across race conditions in either vote (Chi-square for defendant race × final verdict for white men=11.91, p=0.001).

Figure 2. Verdict Choice, Pre- and Post-Deliberation: White Men vs. Others.

In order to control for the jury unit group effects, we conducted the remainder of our analyses of postdeliberation dependent variables using hierarchical linear mixed models with individual juror nested within jury unit, controlling for individual instruction score and age in our analyses between the groups. First we looked at the impact of defendant race and white male status on juror final verdict, then final verdict certain certainty. The analysis of jurors' verdict, nested within jury, indicates a significant main race-of-defendant effect (t(170.0)=2.08, p<0.05), main white male juror effect (t(446.1)=3.08, p<0.01), and race of defendant × juror race/gender interaction effect (t(443.4)=−2.07, p<0.05). Similarly, there were significant effects when using verdict certainty as the dependent measure (race of defendant t(172.6)=−2.81, p<0.01, white male juror t(447.1)=−3.82, p<0.001, and race of defendant × juror race/gender t(444.4)=2.47, p<0.05).

When jurors (nested in jury units) were analyzed as two distinct groups (either white male or not), again, it was clear that the race effects for sentence outcomes were driven by the white male participant subgroup. Specifically, the white male jurors demonstrated significant racial bias against the black defendant relative to the white defendant in final verdict (t(79.30)=3.19, p<0.01) and verdict certainty (t(79.2)=−3.59, p=0.001), whereas the nonwhite and female groups demonstrated no within-group differences in sentence outcomes on either measure as a function of defendant race.

This differential in sentence outcome appears to be substantially the product of the deliberation process. Thus, as illustrated in Figure 2, although the mock jurors on the whole moved toward death after deliberation, the shift toward death was most extreme for the white male jurors in the black defendant condition. To measure this within-subject change, we created a variable that measured shift in verdict certainty scores from straw vote to final verdict and found that the white male participants shifted their certainty toward death from the straw vote to the final vote to a significantly greater degree in the black defendant conditions than in the white defendant conditions (t(88.29)=−2.65, p=0.01). Indeed, these participants moved very slightly toward life for the white defendant (mean=−0.06), whereas they moved a full point toward death on a 7-point scale for the black defendant (mean=1.03). The white men's resistance against a death verdict for the white defendant represented the only circumstance under which there was not a shift toward death in the study. There was no significant difference as a function of defendant race for within-group shift in verdict certainty by defendant race for the women and nonwhite participant group. Again, because of the split indicated on the straw ballots, this shift toward death in general, and particularly for the black defendant, cannot be explained by a pre-existing majority for death.

The Defendant's Profile and Juror Characteristics

Unlike the other participants, the white male jurors also viewed the defendant's criminal “profile” differentially as a function of the defendant's race. Although they did not see the black defendant as more dangerous than the white defendant,Footnote ⁹ they did view him as less remorseful for the crime (t(68.85)=2.35, p<0.05), more cold-hearted (t(62.08)=2.95, p<0.01), and, to a marginally significant degree, as more likely to reoffend (t(86.49)=−1.75, p<0.10).

Participants also answered a series of questions about the defendant's possible motivations for committing the capital crime and were provided options from which to select that were either internal or external in attributional nature, as well as ones that were either potentially aggravating or potentially mitigating. As is illustrated in Table 1, white male participants were significantly more likely to agree with internal, aggravating attributions for the black defendant compared to the white defendant, and significantly less likely to agree with several of the mitigating attributions for the black defendant, particularly those central to this case: that the defendant committed the crime due to a traumatic childhood and that he committed the crime because he suffered from mental illness.Footnote ¹⁰ As Table 1 indicates, the nonwhite and women participants generally did not give significantly different attributions based on defendant race.

Table 1. Attributions About Defendant's Behavior by Defendant Race: White Men Participants vs. Other Participants

Notes: N=169 for white men; N=370 for nonwhites and women.

^** p<0.01; ^*p<0.05; ⁺p<0.1, marginally significant, using linear mixed model hierarchical analysis, with juror nested within jury. Lower number indicates more agreement with that attribution on a 4-point scale.

Our white male jurors not only demonstrated within-group differences in their attributions as a function of defendant race, but they also differed from the other participants in attributing causality, but only in the black defendant condition. Specifically, they were significantly more likely to agree that the black defendant liked to harm others (t(257.6)=3.0, p<0.01) and that he was violent by nature (t(267.8)=1.98, p<0.05) when compared to the remainder of the participants. They were also marginally less likely to attribute the black defendant's crime to a traumatic childhood (t(258.2.64)=−1.90, p<0.06) and marginally more likely to attribute it to the defendant's greed (t(258.1)=1.91, p<0.06) and desperation for money (t(259.6)=1.95, p<0.06) than the other jurors. The two groups of participants did not significantly differ from each other in any of their attributions for the white defendant.

Discussion

The jurors in this experimental study were exposed to exactly the same case and evidence—each word of testimony and argument was identical—in each of the conditions that we employed. Yet the white men who participated as jurors in our study appear to have heard and judged a different case as a function of the race of the defendant. When they judged a black defendant—and essentially only then—they diverged significantly from their peers, both in terms of how they constructed the defendant's blameworthiness and motivation, and on whether they believed he deserved to be allowed to continue to live. Thus, the white male jurors in this study judged a black defendant whom they tended to view as driven by the defendant's own depraved predispositions, and as someone whose criminal behavior they were reluctant to interpret as the product of the defendant's dysfunctional and psychologically damaging background.

As a consequence, these jurors gave significantly less weight to the mitigation that had been offered in order to place the defendant's behavior in a broader and more sympathetic context. White male jurors also saw the black defendant as less redeemable, in that he was viewed as more cold-hearted and remorseless, and as someone who would be more likely to re-offend if given the chance. Furthermore, these effects were evident despite the fact that White male participants had significantly higher comprehension of the capital jury sentencing instructions, something our prior research indicated should have, if anything, moderated racially disparate outcomes and suppressed the between group differences (Reference Lynch and HaneyLynch & Haney 2000). In addition, the significance of the race effects we uncovered in this study is likely somewhat muted, relative to the real world. That is, because we matched our defendants on a number of physical appearance dimensions, the punitive effect of racial stereotypicality in appearance, which has been demonstrated in prior research (Reference EberhardtEberhardt et al. 2006), is suppressed by our methodological design.

Especially problematic, and contrary to our expectations—in light of the mandate to seek consensus in a jury deliberation setting—was that the divergence between the way black and white defendants were perceived and treated grew wider after the deliberation process. By the final verdict, the white male participants stood apart in their extreme punitiveness toward the black defendant whose sentence they decided. Conversely, when considering the white defendant's fate, they were remarkably similar to the other participants, especially once they had deliberated with them. Thus, the race-based punitive tendencies of the white male jurors in this study were amplified rather than muted by the presence of others and the opportunity to publicly discuss their views. This seems to indicate that a form of aversive racism (Reference Gaertner, Dovidio, Dovidio and GaertnerGaertner & Dovidio 1986) may have been activated by the deliberation process itself. It also indirectly indicates that many white male jurors were resistant to alternative perspectives or points of view when they were on juries where the fate of a black defendant was being decided. Moreover, it appeared that these jurors in particular had disproportionate influence on the others during deliberations, moving the other jurors more toward death overall, especially in the black defendant condition.

The racial differences in the evaluation of mitigating evidence (and the importance of these differentials for death sentencing) seem to stem from the fact that white male participants were less able or inclined to empathize with—or strive to understand the plight of—the black defendant. This represents an exacerbation of an already sizable “empathic divide” between the typical juror and defendant in a capital case (Reference HaneyHaney 2004). Yet the jury's genuine understanding of a capital defendant's background and social history and their willingness to grasp that evidence's mitigating significance is what determines whether they will choose life over death (Reference HaneyHaney 1995, Reference Haney2008).

In this sense, the present research represents an initial attempt to map the size and dimensions of this problematic empathic divide, especially the various ways in which race appears to increase its size and effect. Consistent with Reference BowersBowers et al.'s (2001) work, we found that those capital juries that were composed of a significant number (i.e., a third or more) of white men were particularly resistant to the defense penalty-phase case that was presented on behalf of a black defendant. Moreover, this resistance was both amplified and solidified by the deliberation process, where the white male jurors also appeared especially influential in bringing other jurors to their punitive point of view.

These findings appear to have important implications for our understanding of white prejudice against black defendants, and the daunting challenge of creating racial fairness in these most serious and consequential criminal cases. This research reinforces previous conclusions reached by a number of scholars about the way that the race of the defendant shapes criminal justice outcomes. As they have suggested, examining only global outcomes—across decision makers and/or across types of crimes, case characteristics, and offender categories—often masks the multiple ways that race interacts with individual, situational, and/or structural factors to produce and perpetuate inequality (see, e.g., Reference Bushway and PiehlBushway & Piehl 2001; Reference MitchellMitchell 2005; Reference SommersSommers 2007). In the present study, we were able to tease out and analyze the key interactions between juror demographics and defendant race in how the case elements were conceived, weighed, and decided upon, thereby adding to a more contextualized understanding of how and why a defendant's race matters in some cases but not in others.

In addition, this experimental study adds methodological diversity to existing scholarship on race and the death penalty. Our findings converge with studies that employ other methodological designs to examine race effects in capital cases, including quantitative and qualitative analyses of archival and interview data (e.g., Reference Baldus, Woodworth and ZuckermanBaldus, Woodworth, Zuckerman, et al. 1998; Reference BowersBowers et al. 2001). Notwithstanding the sacrifice in mundane realism that jury simulations almost always entail, experimental data such as these add considerable explanatory power to the proposition that race matters greatly in these contexts.

Taken together, all these studies raise important questions about whether and how the death sentencing process can be made racially fair. The death penalty is not meted out in a race-neutral manner, nor is it imposed on persons purely on the basis of their moral culpability. Instead, a variety of impermissible but potent “extralegal” factors (including defendants' and victims' class, gender, and race; local politics; and jurisdictional resources and quality of representation) appear to shape the ultimate outcome of a capital case. Beyond that, other aspects of the overall system of death sentencing compound these problems, serving to exacerbate rather than ameliorate the capital jury's problematic decisionmaking tendencies (Reference HaneyHaney 2005). For example, the death qualification process increases the likelihood that biased jurors will be seated and racially discriminatory sentences produced. Moreover, it appears that the race-based decisionmaking that occurs in capital cases is not simply a product either of individuals or the situations in which they act. Person-situation interactions—demographically related tendencies and the demands of the setting in which these morally complex judgments are made—combine to produce these pernicious outcomes. What remains to be answered is why this situational context, at least as demonstrated by this study, particularly exacerbates racial bias among white men, and not others.

More broadly, these findings complicate social-psychological theorizing about the nature and mechanisms of contemporary racism in that they suggest an interactive process by which racial bias plays out in sociolegal judgment settings. Thus, an integrated theory that takes into account three levels of contributory factors—individual demographic factors, group composition factors, and the influence of decisionmaking processes—may be warranted to understand the conditions under which racism is especially likely to be influential (see Diagram 1 for an illustration of the capital penalty trial scenario). Attending to multiple levels of factors, and considering their interactions, has the potential to be fruitful in explaining why racially disparate outcomes in capital cases persist even in situations where there is no overt racism present, and where individual-level pre-deliberation responses may not necessarily predict racially biased outcomes. As such, these findings converge with recent work by Reference WardWard et al. (2009) indicating the value of looking at how group-level diversity within a criminal justice setting has a significant and independent impact on equitable justice outcomes.

Diagram 1

In addition, these findings strengthen support for theories that highlight differences in positive emotional and cognitive responses in explaining contemporary racism. Indeed, it appears that the majority of the racial disparity is due to a differential assessment of the sympathetic and positive aspects of the defendant and varying levels of empathy for the defendant by our white male participants. These participants not only were less generous in considering mitigating evidence than the others for the Black defendant, but they were also significantly more generous in this regard for the white defendant. Yet because these effects were limited to the white male participant subgroup, such perceptual and affective responses must be understood as at least partly filtered through a specifically raced and gendered worldview or life experience.

Finally, our findings lend further support to the notion that situational and procedural factors matter in the activation of racial animus, as Reference Gaertner, Dovidio, Dovidio and GaertnerGaertner and Dovidio's concept of “aversive racism” postulates (1986). In this case, the deliberation process appeared to trigger the activation of such racism, resulting in a significant shift toward a death verdict for the black defendant and a resistance to such a shift among white men when considering the fate of the white defendant.Footnote ¹¹ Yet contrary to Gaertner and Dovidio's hypothesizing, this dynamic did not cut across all white participants; it was, again, isolated among the white men.

Conclusion

Although the U.S. Supreme Court has recently affirmed and solidified its commitment to ensuring that blacks are not excluded from participation on capital juries by prosecutors who use racially targeted peremptory challenges (Miller-El v. Dretke 2005; Snyder v. Louisiana 2008), issues of genuine racial fairness are much broader, more complicated, and seemingly elusive. As we have noted, for example, a concentration of white male jurors sitting in a black defendant's penalty trial appears to undermine the jury's willingness to consider the case for mitigation. Consequently, ensuring some representation for nonwhite capital jurors may not be sufficient to deliver a truly fair, diverse, and impartial jury. Because the process of minority exclusion is a cumulative one—resulting from the way in which venire pools are constructed, jury eligibility is determined, and death qualification and jury selection are conducted—there is still a high probability that the juries that are finally impaneled to hear and decide a capital case will be susceptible to the kind of white male juror bias that we have uncovered in this study.

There are some additional interventions that could be instituted to mitigate such bias. Thus, particularly in capital cases, where racially diverse juries are especially difficult to seat, courts could adopt policies of “affirmative jury selection” (e.g., Reference FukuraiFukurai et al. 1993; Reference Sommers and Michael NortonSommers & Norton 2008) whereby racial minorities are oversampled in the pool to ensure more substantial representation in the venire and a greater likelihood of significant representation on the jury itself. Another potential strategy might be to use targeted voir dire, designed to make the problem of racial bias salient to potential jurors in a manner that would inhibit the activation of aversive racism (e.g., Reference Sommers and EllsworthSommers & Ellsworth 2001). In addition, strategies may be utilized to more intensely probe white male potential jurors' attitudes and viewpoints during death qualification in cases involving nonwhite defendants to uncover individual-level biases. Finally, any efforts to simplify and clarify the guidelines given to capital juries have the potential to mitigate some racial bias, although it appears that incomprehension has less explanatory value for bias after deliberation (Reference Lynch and HaneyLynch & Haney 2009).

Because white male jurors are more likely to be the beneficiaries of structural privilege and less likely to have personal experience with oppression than their nonwhite and female counterparts, additional steps may need to be taken to present them with persuasive and effective mitigating evidence. That is, even greater effort and special insight may need to be brought to bear in capital cases that involve black defendants, especially if there is a cross-racial dynamic at issue in the case itself. Thus, it seems especially important for defense attorneys, investigators, and experts to recognize the nature of the challenge they confront in such cases, and to meet it through the careful development, thoughtful analysis, and effective presentation of the compelling mitigating narrative that explains the client's life in social contextual terms.

However, we concede that implementing the full panoply of these and other creative strategies may not be enough to entirely overcome racially unfair death sentencing in the United States. For many of the reasons we have discussed here, the post-Furman v. Georgia (1972) reforms that were intended to address standardless decisionmaking on the part of capital juries have not done so. Furthermore, tinkering with the final stage of a process that disproportionately impacts poor and minority defendants at each prior stage will only have limited impact on racially disparate outcomes. To the extent that “biographical racism” (Reference HaneyHaney 2004) contributes to violence; that death-eligible offenses continue to be defined in ways that target poorer, minority defendants; that prosecutors use their considerable discretion to seek death in ways that overvalue white victims and overpenalize nonwhite offenders; and that defendants in the most death-prone jurisdictions often have the poorest quality representation, bias and discrimination will continue to shape the nation's system of capital punishment.