dissenting.
The Court now decides that the death sentence imposed on a young African-American man is proportionate, and that he may therefore be executed.
In settling, once again, the fate of a capital defendant, the Court returns to the difficult issues that surround the proportionality review of a death sentence. Those issues are virtually intractable. Their insolubility reflects not a lack of will, energy, or acuity on the Court’s part. Rather, it stems from the impossibility of evaluating the proportionality of a death sentence that is imposed under a capital punishment regime that is itself founded on conflicting and contradictory principles and is administered without any degree of consistency, much less uniformity.
The hopelessness of proportionality review is exemplified by this case. The Court’s formulation and application of the standards for proportionality remain markedly vague and unworkable. The most serious deficiencies in the Court’s proportionality review are evident, specifically, in (1) the continued use of a universe of cases that includes cases in which the death sentence has later been reversed; (2) the inherent subjectivity of the Court’s principal methods for determining proportionality, frequency analysis, and precedent-seeking analysis and their arbitrary application; and (3) clearly the most significant and least tolerable defect in the Court’s efforts today, the Court’s failure to recognize the recurring indications that the imposition of the death penalty in New Jersey may be infected by racial bias. Those deficiencies of design and application deprive proportionality review of whatever *398faint chance it might have had to provide constitutional legitimacy to the imposition of a death sentence.
I
On April 2, 1983, Cheryl Alston was sexually assaulted and murdered. Defendant, Marko Bey, was subsequently charged ■with and convicted of capital murder, felony murder, aggravated assault, and aggravated sexual assault. See State v. Bey, 112 N.J. 45, 51-52, 548 A.2d 846 (1988) (Bey I). On defendant’s appeal from his sentence and convictions for the Alston murder, this Court overturned both. Id. at 51, 548 A.2d 846. After a second trial on the Alston killing, a jury convicted defendant of purposeful murder. Because defendant, at the time of Alston’s murder, was seventeen-years old, he was deemed not subject to the death penalty and was therefore sentenced to life imprisonment with a thirty-year parole disqualifier for the murder.
About three weeks after the murder of Cheryl Alston, on April 26, 1983, Carol Peniston was sexually assaulted and murdered. Defendant was taken into custody for the murder of Peniston on May 6,1983. After a jury trial, defendant was convicted of capital murder, felony murder, kidnapping, aggravated assault, aggravated sexual assault, robbery, and theft. Id. at 133-34, 548 A.2d 846. A jury sentenced defendant to death for the murder of Peniston. On appeal, this Court affirmed defendant’s convictions, but reversed and remanded for a new trial on sentencing.
The retrial of the penalty phase lasted seven days. The jury unanimously found that the aggravating factors outweighed the mitigating factors, and the jury sentenced defendant to death. Ibid.
Defendant appealed the second sentence of death for the Peniston murder. State v. Bey, 129 N.J. 557, 568, 610 A.2d 814 (1992). Defendant raised objections to the fairness of the second penalty proceeding. He argued that the trial court had prevented him from developing or the jury from fully considering certain mitigating factors. These asserted errors included: prohibiting defense *399counsel from asking leading questions of defendant’s mother, who resisted providing detailed or explicit accounts of her abuse of her son, id. at 593-94, 610 A.2d 814, and excluding the report of a State expert, Dr. Cooke, that largely corroborated the defense theory that defendant, in murdering Peniston, had acted out of aggression toward women, id. at 586-93, 610 A.2d 814.
Over dissents by two of its members, this Court rejected those contentions, either determining that no error existed or finding that although error was present, it was “extremely unlikely” that the error “had the capacity to affect the jury’s deliberations or produce an unjust result.” 129 N.J. at 616, 610 A.2d 814.
II
The Capital Punishment Act provides that at a defendant’s request, “the Supreme Court shall determine whether the sentence is disproportionate to the penalty imposed in similar cases, considering both the crime and the defendant.” N.J.S.A. 2C:11-3e. Although not required by the federal constitution, Pulley v. Harris, 465 U.S. 37, 104 S.Ct. 871, 79 L.Ed.2d 29 (1984), proportionality review is intended as a safeguard against the arbitrary and capricious imposition of the death penalty.
This Court has acknowledged the “entirely unique function” of proportionality review in a capital proceeding. State v. Ramseur, 106 N.J. 123, 326, 524 A.2d 188 (1987). The Court in Ramseur held that because death is fundamentally different from all other punishments, a correspondingly greater need for reliability existed in determining whether death was appropriate in a given case. Ibid. Thus a practical and effective proportionality review is essential to the Court’s continued insistence that the imposition of the death penalty can be carried out in fidelity to the constitutional norms that bind state power.
Proportionality review in New Jersey is intended to be offender rather than offense-oriented. The Court in State v. Marshall, 130 N.J. 109, 613 A.2d 1059 (1992) (Marshall II), explained that the offender-oriented approach inquires whether the penalty in a *400particular case is disproportionate to the punishment imposed on others convicted of the same crime. The touchstone of the Court’s analysis is ensuring that the death penalty is “imposed fairly, and with reasonable consistency.” Ibid.
In Marshall II, the Court concluded that proportionality review required a determination of deathworthiness that was informed as fully as possible. That determination called for the creation of a “universe” of cases sufficiently inclusive to allow a broad-based consideration of all relevant factors bearing on the deathworthiness of a homicide defendant, including both jury and prosecutorial decisions. Id. at 137, 613 A.2d 1059.
The Court then settled on two complementary means of assessing the proportionality of the defendant’s death sentence in relation to various sub-groupings of cases deemed to be similar to the defendant’s. The first approach, generally referred to as “frequency analysis,” is composed of three discrete statistical methodologies each intended to give the Court a different perspective on the proportionality of the defendant’s sentence relative to other similar cases. The controlling evaluative focus of the statistical approach is on the relative frequency with which a death sentence is imposed in cases deemed to be similar. Id. at 153, 613 A.2d 1059. The second approach authorized by this Court in Marshall II, “precedent-seeking analysis,” relies on the statistical analysis as a point of departure. The precedent-seeking approach allows the Court, on a case-by-case basis, to compare similar eases according to a defendant’s criminal culpability. Id. at 155, 613 A.2d 1059.
The statistical and precedent-seeking approaches are intended to complement one another. Theoretically, they afford the Court a perspective informed both by empirical analysis and the more traditional manner of case-by-case, judicial assessment. That dual approach, combining statistical and precedential analyses, remains at the core of what the Court does in carrying out proportionality review.
*401Proportionality review is also designed to provide a window through which the Court can monitor the administration of the death penalty in New Jersey. See Ramseur, supra, 106 N.J. at 327, 524 A.2d 188. In Marshall II, the Court articulated its strong concern about the influence that invidious factors, particularly race, might exert on the administration of the death penalty. 130 N.J. at 207-08, 613 A.2d 1059. Proportionality review attempts to respond to that concern. In this regard it should be noted, however, that the Court’s authority and duty to guard against the presence of certain invidious factors, such as racial discrimination, in the administration of the death penalty do not flow from the Capital Punishment Act. The Equal Protection Clause of the Fourteenth Amendment and New Jersey’s state constitution prohibit state action that discriminates on the basis of, inter alia, a person’s race. Were a defendant to allege racial discrimination as a determinant of the decision to prosecute her for or convict her of a capital crime, or to sentence her to death, this Court — independent of any statutory or constitutional mandate to conduct proportionality review — would be constitutionally required to examine meticulously that allegation. The method of that examination would be very similar to, if not functionally indistinguishable from, what we now call “proportionality review.”
Conceptually, the methodology of proportionality review aspires to fulfill two concurrent purposes: (1) to ensure that a particular death sentence is not disproportionate; and (2) to ensure that invidious factors, especially race, are not at work in determining who receives death sentences. The failure of the Court’s methodology to achieve those purposes is demonstrated in today’s decision.
Ill
An issue of critical importance in the determination of capital-sentencing proportionality is the definition of the appropriate universe of eases from which to conduct judicial review. The bounds of that universe are effectively dictated by the objectives *402of proportionality review. Because the Court has recognized that proportionality review should serve to prevent “any impermissible discrimination in imposing the death penalty,” Ramseur, supra, 106 N.J. at 327, 524 A.2d 188, the race, sex, and socioeconomic status of the defendants, as well as the role of geography, by county, in sentencing decisions are factored into the definition of the universe. Further, because prosecutorial as well as jury decisions about deathworthiness are relevant to assessing whether invidious discrimination is at work, the Court concluded that the outer limits of that universe should include all death-eligible homicides whether or not prosecutors have chosen to prosecute them as capital crimes.1
One enduring and troubling issue emanating from Marshall II is the Court’s decision to include in its universe capital cases later reversed on appeal as death-sentence cases for purposes of proportionality review. Marshall II does not provide any extended explanation or defense of the Court’s choice to treat reversed cases as death-sentenced. The Special Master, in dealing with that problem, constructed three possible responses: (1) treat all reversed cases as valid death sentences, (2) treat no reversed case as a valid death sentence, (3) proceed on a case-by-ease basis, assessing the reliability of the original sentence for use in propor*403tionality review. David Baldus, Death Penalty Proportionality Review Project, Final Report to the New Jersey Supreme Court, 61 (Sept. 24, 1991) (hereinafter Final Report). The Special Master recommended the third approach to the Court, noting that a presumption should exist against using death-sentences reversed because of penalty-phase errors. Id. at 63.
The Court today attempts a defense of the use of reversed death sentences in proportionality review. It does so using evasive language, noting that “[i]n the absence of an acceptable explanation, we continue to believe that a death sentence, even when reversed, represents a societal consensus concerning the deathworthiness of a defendant.” Ante at 348, 645 A.2d at 692 (emphasis added). The Court cites Marshall II, arguing that “issues pertaining to procedural fairness, as distinguished from those that affect the substance of a crime, do not bear on a jury’s determination of deathworthiness.” Ante at 347, 645 A.2d at 691 (emphasis added). Yet nowhere does the Court define what an “acceptable explanation” might be. Nor does it illuminate or even dwell on the obvious over-simplification that “issues pertaining to procedural fairness * * * do not bear on a jury’s determination of deathworthiness.” Ibid.
The Court also notes that after reversal, prosecutors do not always re-try cases as capital crimes, observing that the reasons for that failure are “varied and indeterminable.” Ante at 348, 645 A.2d at 692 (emphasis added). Nevertheless, the Court is content to draw the inference that “we cannot conclude that in any given case that a life sentence resulted from the view that the defendant was not initially deathworthy, rather than, for example, the strength of the prosecutor’s case.” Ibid. That is a curious statement, and the Court does not bother to explain what it means by the “strength of the prosecutor’s case.”
I strongly criticized those same basic conclusions as they were expressed in Marshall II, supra, 130 N.J. at 253-57, 613 A.2d 1059 (Handler, J., dissenting), and I find myself no more persuaded by the Court’s explanations today. Reversed death sentences *404should be considered life sentences for the purpose of proportionality review. Id. at 253, 613 A.2d 1059. My opinion is based on the premise that reversed death verdicts are, by definition, unreliable indicators of deathworthiness.
The irrationality of using reversed death sentences in proportionality review should be obvious. Yet the Court persists in indulging what I have termed an “unfathomable irony” whereby a reversed sentence, by definition too unreliable for use in sentencing a defendant, is nonetheless used collaterally against yet another defendant on proportionality review. Ibid. The infirmity of the Court’s reasoning stems from its insistence that a defendant’s “deathworthiness” is somehow distinguishable from the final legitimate verdict reached in that defendant’s case.
To justify the use of reversed death sentences in proportionality review, the Court draws a facile but radically unsound distinction between “procedural fairness” and “the substance of the crime.” Ante at 347, 645 A.2d at 691. The Court intimates that a jury’s determination of a defendant’s deathworthiness is somehow separable from the rules that structure the jury’s deliberative process. Nothing could be more at odds with the soul of capital-murder jurisprudence, as it has evolved in the post-Furman era. The moral logic, such as it is, behind the continued constitutionality of the death penalty is the belief that deliberative procedures can be structured reasonably to ensure that the death penalty is not imposed arbitrarily, i.e., that only those who are truly “deathworthy” receive death sentences. See Ramseur, 106 N.J. at 183, 524 A.2d 188 (commenting on Furman v. Georgia, 408 U.S. 238, 92 S.Ct. 2726, 33 L.Ed.2d 346 (1972)). That logic insists that both the procedural structure and substantive quality of a jury’s deliberations are integral to the ultimate constitutionality of a death sentence. Realistically, procedural fairness and substantive culpability are interwoven throughout the prosecution of a capital case; the sentencing process starts with the qualification of the very first juror.
*405Even the most cursory review of our case law since Ramseur indicates our efforts to refine our capital-murder doctrine to ensure that only the truly deathworthy are able to receive a death sentence. See State v. Clausell, 121 N.J. 298, 345-46, 580 A.2d 221 (1990) (requiring that juries be informed of legitimacy and acceptability of non-unanimous, non-death-deserving verdict at the penalty trial); State v. Gerald, 113 N.J. 40, 85, 549 A.2d 792 (1988) (requiring that to be death-eligible, defendant must have intended to kill, not merely to have inflicted serious bodily injury); State v. Williams, 113 N.J. 393, 453-54, 550 A.2d 1172 (1988) (reversing death sentence because prosecution relied on victim-impact evidence); State v. Bey, 112 N.J. 123, 162-77, 548 A.2d 887 (1988) (requiring reversal if Court incorrectly instructed jury concerning finding and weighing of mitigating factors); Ramseur, supra, 106 N.J. at 211, 524 A.2d 188 (narrowing grounds for c(4)(c) aggravating factor to infliction of severe suffering). Those decisions reflect more than mere procedural tinkerings. They reflect continuing consideration of the factors that appropriately determine deathworthiness.
The Court misunderstands the unbreakable link between the structure that frames the jury’s deliberative process and the substantive standards that inform its ultimate determination of deathworthiness. Its crude distinction between substance and procedure erodes the very foundation of the death penalty’s constitutionality — that is, the belief that truly deathworthy defendants can be identified by juries if they are given a deliberative structure within which their discretion is “suitably directed and limited so as to minimize the risk of wholly arbitrary and capricious action.” Gregg v. Georgia, 428 U.S. 153, 189, 96 S.Ct. 2909, 2932, 49 L.Ed.2d 859, 883 (1976) (Stewart, Powell, and Stevens, JJ.). Death sentences are reversed because at some point the carefully constructed environment of a capital-punishment prosecution has been contaminated, whether by procedural or substantive fault, thereby undermining the soundness and impugning the reliability of the jury’s ultimate determination of deathworthiness.
*406The Court “acknowledges”, as, at a minimum, it must, that a reversed death sentence is “a less persuasive indicator of death-worthiness” than an affirmed death sentence. Ante at 348, 645 A.2d at 692. It leaves us totally unenlightened, however, as to why such a problematic indicator of death worthiness should be used to justify the death sentence of another defendant.
At the very least, the Court should impose a rebuttable presumption that reversed death sentences are invalid determinations of deathworthiness. That approach was endorsed by the Special Master. He expressed the opinion that some reversed death sentences might be salvageable for use in proportionality review especially if, on re-trial, another death sentence was imposed. Final Report at 61-62. The Special Master suggested that the Court analyze each reversed sentence proposed for use in proportionality review to determine if the error requiring reversal impugned the reliability of the original sentence. Ibid. Yet here the Court inverts the structure of proof inherent in the Special Master’s proposal. The Court states that “defendant does not explain why some errors that have caused us to reverse the death sentence necessarily reflect on the jury’s ability to assess the defendant’s deathworthiness.” Ante at 347-348, 645 A.2d at 691-692. With due respect, at a bare constitutional minimum, the Court should explain and the State should bear the burden of demonstrating why errors requiring reversal of a death sentence do not necessarily reflect on deathworthiness.
The Court itself concedes the dramatic impact that excluding reversed death sentences would have on its exercise of proportionality review. Ante at 346, 645 A.2d at 691. Of the seven other prior-murder capital cases in the universe, all seven were reversed for errors of one kind or another, leaving Marko Bey as the only prior murderer to be sentenced to death.
The Court peremptorily reaches out to note that the fact that Marko Bey is the only prior murderer finally sentenced to death would not compel a finding that his sentence was disproportionate. Ibid. Whatever the merits of that hypothetical conclusion, that *407case is not before the Court. Because the Court persists in following a course that contradicts the basic principles of capital-murder doctrine and common sense, Bey is condemned by a proportionality review that uses as its primary feature comparison to seven other cases in which verdicts of death have been rejected by this Court as unreliable. The Court’s continued use of reversed death sentences dooms its proportionality review from the start.
IV
The Court’s conclusion that defendant’s death sentence is proportionate is the end of a process of review that is itself incoherent. The inherent subjectivity of proportionality review, to which the Court so readily admits, ante at 345, 348, 363, 369, 645 A.2d at 690, 692, 699, 702, is exacerbated by the Court’s flaccid treatment of the grave methodological problems apparent in both its frequency analysis and precedent-seeking review.
A.
The Court’s application of frequency analysis reveals a palpable bias favoring the proportionality of a death sentence. The Court engages in that analysis without a settled standard or perception of what constitutes a high or low predicted frequency of death. Consequently, it inevitably ends up engaging in a form of reasoning that is little more than a selective and convenient rationalization of proportionality.
Theoretically, frequency analysis operates according to a basic controlling principle, namely, that “ ‘[a] death sentence is comparatively excessive if other defendants with similar characteristics generally receive sentences other than death for committing factually similar offenses.’ ” Marshall II, supra, 130 N.J. at 153-54, 613 A.2d 1059 (quoting Tichnell v. State, 297 Md. 432, 468 A.2d 1, 17 n. 18 (1983)) (alterations in original). The higher the frequency of a death sentence among a comparison group, the more certain the determination that the sentence is proportionate.
*408The Court has understandably refused to employ a precise numerical cut-off in determining proportionality. Yet it refuses to state, even in general terms, what a high or low frequency might be. The Court contents itself with what it terms a “general standard.” Ante at 351, 645 A.2d at 693. The Court rests secure in the notion that “a standard that applies generally is the antithesis of one that applies arbitrarily.” Ibid. But what does “generally” mean? What could it reasonably mean? In common parlance, if something generally occurs then it happens more often than not. That would imply at least a fifty percent threshold. Yet the court quite clearly states that “generally” does not require a threshold rate over fifty percent. Ante at 351, 645 A.2d at 693.
The conclusion to be drawn from the Court’s use of a “general standard” is that it simply does not have or is unwilling to announce a pre-existing standard for determining high or low frequencies of death sentence imposition. The Court’s refusal to recognize its own ambivalence only serves to deepen the confusion surrounding its application of frequency analysis with the result that the Court’s assessments based on frequency analysis appear perfunctory and wholly unconvincing.
The absence of a clear standard invites manipulative reasoning that inevitably tilts toward the conclusion that a sentence is not disproportionate. That is especially so when, as is the case here, frequency analysis is given such a subservient role in comparison to precedent-seeking analysis. The Court’s now persistent refusal to clarify the nature of the relationship between frequency analysis and precedent-seeking analysis creates a situation in which the Court has entirely too much leeway to pick and choose among the results of the various tests.
The point is not to suggest that the Court adopt a precise numerical cut-off. Rather, the Court should come to a rough conclusion, before it begins to analyze the statistical data that make up frequency analysis, about which general range of frequencies are low and which are high. If the Court, as here, first *409assesses where defendant falls and then determines if that statistical ranking represents a high or low enough frequency, the sense that the Court is merely using the statistics to vouch for a preconceived, subjectively-comfortable conclusion about proportionality is unavoidable. Obviously, if a defendant’s ranking falls at the extremes (high or low), those concerns are minimal, but in cases, such as this one, in which defendant’s rank falls roughly in the middle, those concerns become pronounced.
A frequency analysis of defendant’s sentence reveals that death-sentencing frequencies for cases similar to his range from less than ten percent to about sixty percent. Given that that amounts to a “toss of the coin,” as defendant argues, the conclusion should be that his death sentence is disproportionate. Minimally, the inconclusive frequency of death-sentencing in cases such as defendant’s bespeaks the susceptibility of such eases to impermissible factors such as racial discrimination. See discussion infra at 362-363, 645 A.2d at 699.
Falling as defendant does in the mid-range range of predicted frequencies of death sentence imposition, he is peculiarly vulnerable to the ad hoc character of this Court’s frequency analysis. The Court’s analytical failure with respect to frequency analysis serves only to further devalue the already scant protections defendant is afforded by frequency analysis. Although originally designed as the more objective of the two methods that make up proportionality review, frequency analysis as applied by the Court does little more than set the stage for whatever subjective determination might be made under the precedent-seeking approach.
B.
The Court has acknowledged that it relies more heavily on precedent-seeking analysis. Unfortunately, as designed by Marshall II, and most especially as practiced by the Court today, precedent-seeking analysis is woefully inadequate to its task.
The ostensible virtue of precedent-seeking analysis is that, unlike a statistically-based frequency analysis, it employs a more *410traditional form of judicial reasoning. The Court assesses the proportionality of a defendant’s death sentence by seeking out patterns of common facts and circumstances that inform a comparative judgment about the relative culpability, and thus the death-worthiness, of a particular defendant. The inherent subjectivity of that approach is undeniable. The Court posits that subjectivity as a welcome antidote to the sterile and deterministic nature of statistical analysis. Thus the Court gamely asserts that during precedent-seeking review, it comes to meet the “real people involved in defendant’s and similar cases.” Marshall II, supra, 130 N.J. at 154, 613 A.2d 1059.
As with the standardless use of frequency-analysis, the Court’s precedent-seeking review is rife with rationalization posing as objectivity. The Court quite reasonably suggests that age — a defendant’s youthfulness — need not axiomatically correspond to a decreased culpability, ante at 360, 645 A.2d at 697-698, although it does concede that age “often mitigates a defendant’s culpability.” Ante at 383, 645 A.2d at 709. The Court does not speculate about why this jury was unimpressed by defendant’s youth. It is content with the proposition that the jury’s failure to find age as a mitigating factor does not make its determination per se aberrant. Ibid.2
Although Bey’s youthfulness in itself, may not have constituted a mitigating factor, the Court inexcusably treats that mitigating factor in a vacuum without the slightest appreciation that separate' mitigating factors, particularly in the weight ascribed to them, *411may be viewed by a jury as interrelated. Thus, defendant’s youthfulness might serve to strengthen the mitigating effect of the evidence of the abusive childhood he had endured and, conversely, the abuse he had experienced as a child could greatly affect the significance to be attributed of his youthfulness.
The Court rightly notes that the experience of childhood abuse is, unfortunately, a trait common to many defendants faced with a death sentence. Ante at 384, 645 A.2d at 709. Indeed, the evidence adduced at Bey’s penalty trial — his mother’s alcohol-abuse, the abandonment of his family by his father, the frequent and savage beatings, the emotional deprivation and consequent alienation, the early dependence on drugs and alcohol and its resultant destruction of mental and social capacities, and finally the maladaptive resort to violence — is a litany all too familiar to any sentencing judge. Yet what distinguishes Bey’s case is his youth at the time of his crimes. He was not yet eighteen at the time he committed them. That his behavior is horrific and deserving of severe punishment is not in question. What ought to be in question, however, and is not, is whether relative to other defendants who have experienced similar abuse, Bey’s youthfulness distinguishes him as less culpable. The Court’s failure to address that question marks yet another in a series of seemingly arbitrary determinations that characterize this Court’s use of precedent-seeking review.
The Court’s shortsightedness with respect to the correlation between defendant’s age and the evidence of the abuse he suffered as a child takes an even more saturnine form in the manner in which the Court distinguishes defendant’s case from that of James Koedatich.
By the Court’s own standards, Koedatich, a prior murderer with a pattern of sexually assaulting his victims, is the ideal comparison case. Ante at 384, 645 A.2d at 709. At pains to differentiate Koedatich, who received a life sentence, the Court quite brazenly asserts that he “unlike Bey, offered additional, uncontroverted *412evidence to demonstrate the impact of the abuse and violence [he had] suffered.” Ante at 384, 645 A.2d at 710.
The Court’s attempts to distinguish Bey based on the fact that Koedatich offered more and better evidence of child abuse expose the radical contradictions that permeate our capital punishment system. More disturbing, they show the Court’s willingness to paper over its own role in weakening the evidence of Bey’s abusive childhood.
Bey argued forcefully on direct review that the trial court erroneously had prevented him from fully adducing evidence related to the abuse he had suffered as a child. The trial court refused to allow into evidence the Cooke Report which, although prepared by a psychologist retained by the prosecution, addressed Bey’s mental and emotional condition at the time of the murders and went to the core of his defense. In particular, the Cooke Report linked the abuse Bey had suffered as a child to his rage and violence against his victims. It thus “demonstrated the impact of the abuse and violence [he had] suffered,” which is precisely what the Court, now conducting proportionality review, says is lacking in the evidence that was adduced with respect to Bey’s childhood. Ante at 384, 645 A.2d at 710. Thus, exclusion of that evidence was, at one point on direct appeal, harmless. Yet, currently, before the same Court conducting proportionality review, the absence of that evidence is a factor that supports the proportionality of a death sentence. What was once harmless is now lethal.
In the same vein, Bey’s lawyers were refused permission to treat Bey’s mother, a defense witness understandably reluctant to give detailed testimony about the abuse she had inflicted on her son, as a hostile witness. The sentencing court’s ruling prohibited defense counsel from asking Ms. Bey leading questions. Blocked, defense counsel was unable to elicit detailed accounts of the manner, severity, and frequency of the abusive incidents in the young Marko Bey’s life. Now the Court affirms the proportionality of Bey’s death sentence in part because of his failure to adduce *413sufficient evidence of abuse that distinguishes him as more culpable than James Koedatich.
Through a series of subjective judgments that ignore the impact of the Court’s own prior harmless-error rulings, the Court deprecates glaring indications of jury aberrancy. Although Bey is the youngest capital defendant yet to face the Court, no juror found the “age” mitigating factor. Although Bey produced substantial, albeit incomplete, evidence of abuse during childhood, only two jurors found the “mental disturbance” mitigating factor. Rather than test those “hypotheses” of aberrancy by comparing Bey’s sentence with those of young defendants or defendants where the mental disturbance factor was found, or even better, where both were found, the Court instead discounts the relevance of those factors and concentrates instead on a comparison of Bey to the narrow class of prior murderers. Then, when finally confronted with Koedatich, Vasquez, and Booker, all prior murderers, who, according to the Court, exhibit similar levels of culpability, ante at 386, 645 A.2d at 710, all of whom received life-sentence, the Court distinguishes them because, as noted, Koedatich offered more and better evidence of child abuse, Vasquez’s parent’s pled for his life, and Booker’s murderous spree was fueled by substance abuse. Ibid. Yet the narrative summaries reveal that both Booker and Koedatich killed three persons, and both were significantly older than Marko Bey, while Vasquez, who strangled a thirteen-year old child to death, denied any mental-health problems. See Appendix, infra at 431, 645 A.2d at 733.
To bolster its attempts to distinguish Koedatich’s case from that of defendant, the Court indulges in a form of argumentation more appropriate to journalism than a judicial decision. Koedatich’s life sentence was the result of a lone hold-out juror. The Court suggests that this juror “prevented” the remaining eleven of the opportunity to return a death verdict implying that somehow Koedatich, within a single juror’s vote of being sentenced to death, is really more deathworthy than not. Ante at 387, 645 A.2d at 711. The Court’s observations and its studied implication border *414on the outrageous. Jury unanimity is essential to a determination of deathworthiness. Ramseur, supra, 106 N.J. at 301, 524 A.2d 188. Neither our statute nor the state or federal constitution permits a classification of defendants who are “almost deathworthy.” Yet is that not precisely what the majority does in its attempts to distinguish James Koedatich?
It is simply wrong for this Court to suggest, let alone rely on and give legal force to, the notion that some non-deathworthy defendants are less deathworthy than others. Such thinking is dangerously misconceived. It is totally at odds with what is a fundamental condition of deathworthiness, namely, the unanimous concurrence of qualified jurors that the defendant’s life shall be forfeited. Resort to such an argument betrays the Court’s enduring confusion about the basic constitutional norms that inform and regulate our capital punishment regime.
The Court’s exercise of precedent-seeking review not only serves as a grim reminder of the Court’s disastrous rulings in Bey TV, but also should impel the Court to consider the viability of the harmless-error doctrine as it is currently practiced in the review of capital cases. We see now quite clearly that error adjudged harmless at one stage of a capital case can materialize at yet another stage exerting substantial, perhaps dispositive, influence on subsequent determinations that bear directly on the imposition of a sentence of death.
C.
The Court’s decision today highlights the impossibility of harmless error analysis in the penalty trial of a capital case. Errors in a defendant’s penalty trial that were adjudged harmless on direct review resurface in proportionality review where their effect can be incalculably harmful. This Court, having earlier determined certain errors to have been harmless because they were “not clearly capable of producing an unjust result,” State v. Bey, 129 N.J. 557, 591, 610 A.2d 814 (1992) (Bey IV), now must confront the *415inescapable reverberations of its failure to foresee the continuing capacity of “harmless error” to contribute to an unjust result.
Two distinct conclusions emerge from a consideration of harmless-error-review undertaken in light of the subsequent proportionality analysis done in this case. The first is that the Court’s prior determination that errors in the penalty trial were harmless in terms of their capacity to affect the sentence imposed on the defendant is inaccurate. The second, less obvious, although no less damning, conclusion to be drawn from the Court’s exercise of proportionality review is that harmless-error analysis, traditionally understood, is not appropriate or feasible in the penalty phase of a capital case.
From the very start, the problems inherent in the application of conventional harmless-error analysis to the penalty phase of a capital case have been evident. The United States Supreme Court first allowed the application of harmless-error analysis to errors made in the penalty phase of a capital case in Satterwhite v. Texas, 486 U.S. 249, 108 S.Ct. 1792, 100 L.Ed.2d 284 (1988). Even as the Supreme Court sanctioned the use of harmless-error review, it felt compelled to note that the “evaluation of the consequences of an error in the sentencing phase of a capital case may be more difficult because of the discretion that is given to the sentencer.” Id. at 258, 108 S.Ct. at 1798, 100 L.Ed.2d at 295 (emphasis added); see also Clemons v. Mississippi, 494 U.S. 738, 754, 110 S.Ct. 1441, 1451, 108 L.Ed.2d 725, 742 (1990) (“In some situations, a state appellate court may conclude that peculiarities in a case make appellate reweighing or harmless-error analysis extremely speculative or impossible.”). Justices Marshall and Brennan, concurring in Satterwhite, argued that harmless-error review was impossible in the penalty phase because unlike a traditional guilt-phase determination, a jury’s decision in the penalty phase is a “profoundly moral evaluation of the defendant’s character and crime.” 486 U.S. at 261, 108 S.Ct. at 1800, 100 L.Ed.2d at 297 (Marshall, J., concurring). Given the complex nature of the jury’s deliberation in the penalty-phase, “predicting *416the reaction of a sentencer ... on the basis of a cold record is a dangerously speculative enterprise.” Ibid.
The conceptual basis of harmless-error review — that a court can determine the effect of the error on the verdict — is not viable in a penalty-phase trial that requires juries to make value determinations rather than simply to find facts. Linda E. Carter, Harmless Error In the Penalty Phase of a Capital Case: Doctrine Misunderstood and Misapplied, 28 Ga.L.Rev. 125, 149 (1993) (Hereinafter “Carter”). “The individual choices jurors make [in the penalty-phase trial] about the existence of mitigating circumstances coupled with the unique weighing of factors creates a proceeding fundamentally different from the guilt trial.” Ibid.
Moreover, traditional harmless-error analysis cannot be fitted to the distinctive nature of a penalty-trial proceeding. The typical harmless-error inquiry is focused on the amount of evidence adduced to support a particular conclusion. Carter, supra, 28 Ga.L.Rev. at 159. Thus, on direct review of defendant’s case, the Court determined that the exclusion of the Cooke Report and the refusal to allow the defense to elicit the testimony of Ms. Bey as a hostile witness were harmless errors, in part, because they constituted cumulative evidence. See Bey IV, supra, 129 N.J. at 590, 610 A.2d 814 (“Viewed in the context of the medical testimony at trial, Dr. Cooke’s report offered compelling but cumulative evidence of the defendant’s background and personality disorder.”) (emphasis added); id. at 594, 610 A.2d 814 (“Moreover, Mrs. Bey’s testimony was cumulative[.]”). Such a quantitative review does not, indeed cannot, take into account the complex value judgments that constitute juror determinations in the penalty-phase.3 In the penalty phase of a capital case, where the sentencer’s discretion contemplates the influence of moral values, see McCleskey v. Kemp, 481 U.S. 279, 294, 107 S.Ct. 1756, 1767, 95 L.Ed.2d 262, *417279-80 (1987) (“Each jury is unique in its composition, and the Constitution requires that its decision rest on innumerable factors.”), the Court must consider infinitely more than what a “reasonable” juror might do. See Comment, Deadly Mistakes: Harmless Error in Capital Sentencing, 54 U.Chi.L.Rev. 740, 756 (1984).
A defendant’s right to proportionality review further debilitates the prospects that a court can ever fairly apply conventional harmless-error analysis to errors occurring in the penalty phase of a capital case. We have acknowledged the potential prejudice that can result from the fact that the evidence adduced (or excluded) at one phase of a capital-murder prosecution will overlap another phase of the prosecution. State v. Erazo, 126 N.J. 112, 133, 594 A.2d 232 (1991) (noting that evidence introduced at guilt phase of capital trial has an inescapable impact on jury’s deliberation in penalty phase). In most cases, the prosecution moves guilt-phase evidence into the record for the penalty trial. But even if the State does not explicitly incorporate guilt-phase evidence into the penalty-trial, “the danger abides that the jury will rely on it during the penalty-phase deliberations.” Ibid. The flow of evidence from the guilt to the penalty phases of a capital trial is an inescapable problem in any system, that uses, as we do, a bifurcated trial structure but sits the same jury for both the guilt and the penalty phases. By virtue of the Court’s decision today, that evidence can now slide over into proportionality review. Ante at 368, 645 A2d at 701. Thus errors made at the guilt and penalty phases are swept along into proportionality review.
My point is not that the Court is necessarily wrong in limiting its precedent-seeking review to the evidence before the jury that actually rendered the death sentence, but rather that in accepting the evidence on the record before the original jury, the Court has an obligation to recognize that it accepts whatever errors (of inclusion or exclusion) might have been made in assembling that body of evidence. Moreover, the Court must recognize that its prior harmless-error rulings cannot properly account for the effect *418those errors might have on the subsequent determination of the defendant’s sentence’s proportionality.
In my view, a court that on proportionality review chooses to limit its review to the “evidence” before the jury must be willing to re-consider its prior harmless-error determinations of penalty-phase error to account for such effect that those errors may have on the proportionality determination. Further, a court must conclude that given the complex value determinations required of a jury at the penalty phase and given the impossibility of predicting the likely effects of evidentiary errors on proportionality review, conventional harmless-error analysis of penalty-trial errors is unworkable.
The values that lay behind the harmless-error doctrine — finality, conservation of judicial resources, and the determination not to allow irrelevant error to compromise the integrity of the judicial process — are substantial, but by force of reason and morality, they are dwarfed by the “awesome severity of a death sentence.” Satterwhite, supra, 486 U.S. at 264, 108 S.Ct. at 1801, 100 L.Ed.2d at 298 (Marshall, J., concurring). We see in today’s decision the insidious effects of trial errors sanctioned by a harmless-error analysis that is ill-fitted to the nature and purposes of a penalty trial and cannot begin to anticipate its implications for a subsequent proportionality review.
V
The role that comparative proportionality review plays in preventing invidious discrimination is perhaps its most significant one. The United States Supreme Court has recognized that when state procedures provide adequate protections in the prosecution of a capital case, proportionality review is not required by the federal constitution. Pulley, supra, 465 U.S. at 45, 104 S.Ct. at 876, 79 L.Ed.2d at 37. However, despite the lack of a federal constitutional requirement, our death-penalty statute affords capital defendants the right to seek proportionality review of their sentences. N.J.S.A 2C:ll-3(e). Indeed, when the Legislature *419was considering the present capital-sentencing scheme in 1982, the Attorney General himself recognized the importance of guarding against invidious discrimination in sentencing defendants to death, urging the Legislature “to make sure that [death] sentences are being meted out in a fair, even-handed way throughout the State, and that we do not' have either classes of individuals or areas in the State which appear to be arbitrary one way or other.” Joseph H. Rodriguez, Michael L. Perlin & John M. Apicella, Proportionality Review in New Jersey: An Indispensable Safeguard in the Capital Sentencing Process, 15 Rutgers L.J. 399, 429 n. 203 (1984).
The Court has yet to confront the question whether proportionality review is required under our State Constitution. That, I submit, is a question of overarching magnitude the answer to which cannot be long postponed. Yet, as noted, supra, at 7, even without a constitutional mandate to exercise proportionality review, the Court has an independent obligation to review the capital-punishment system to ensure that it is not infected by invidious discriminatory factors. The Court has stated clearly that “ ‘[discrimination on the basis of race, sex, or other suspect characteristics cannot be tolerated.’ ” Marshall II, supra, 130 N.J. at 135, 613 A.2d 1059 (quoting Ramseur, supra, 106 N.J. at 330, 524 A.2d 188) (alteration in original). With regard to race, the Court has stated unequivocally:
We have committed ourselves to determining whether racial and ethnic bias exist in our judicial system and to “recommend ways of eliminating it wherever it is found." Hence, were we to believe that the race of the victim and race of the defendant played a significant part in capital-sentencing decisions in New Jersey, we would seek corrective measures, and if that failed, we could not, consistent with our State’s policy, tolerate discrimination that threatened the foundation of our system of law.
[Id. at 209, 613 A.2d 1059 (citations omitted) (emphasis added).]
Charges of racial bias within our capital-sentencing system are not new. The Special Master’s Report, noted in Marshall II, suggested that a discrepancy in capital-sentencing rates may correlate to the race of the defendant or the race of the victim. 130 N.J. at 207, 613 A.2d 1059. The Court, in Marshall II, however, rejected the defendant’s race discrimination claims be*420cause they were not “relentlessly document[ed].” Id. at 213, 613 A.2d 1059. According to the Court, the disparities were not consistently shown — that is, the data that indicated race-of-defendant disparities did not show race-of-victim effects in the penalty-trial decisions, and the race-of-victim effects in cases advancing to trial were less stable than the effects observed for the race-of-defendant variable in the penalty trial decisions. Id. at 212-13, 613 A.2d 1059. This Court announced that it was not yet convinced that the effects of racial discrimination were systemic. Id. at 213, 613 A.2d 1059. Clearly, however, the Court was put on notice of preliminary indications that racial discrimination might well be at work in determining the rates at which certain cases were charged as capital crimes, and in determining the rates at which death sentences were actually imposed.
According to the Court in Marshall II, the Public Defender’s most compelling argument was that for cases within the mid-range of aggravation, an African-American had a sixty-four percent greater risk of being sentenced to death. The Court, however, accepted the Special Master’s conclusion that those results were not conclusive. The Special Master advised that “more work will be required to determine if they persist under closer scrutiny and alternative analyses.” Final Report, supra, at 101. The Court concluded by stating that the statistical indices of discrimination were “not sufficiently alarming to compel a conclusion of substantial discriminatory effect.” Marshall II, supra, 130 N.J. at 212, 613 A.2d 1059.
Defendant now claims that the tentative findings, noted but found inconclusive in Marshall II, have now been verified by further analysis. In my view, the Court cannot disregard or discount the evidence that defendant has produced demonstrating the heightened risk of death sentencing for African-American defendants, particularly at the mid-range levels of culpability.
Defendant bases his arguments on an updated version of the data base used by the Special Master in Marshall II. Forty cases have been added to the original data set. In addition, defendant’s *421experts, Messrs. Weiner and Mills of Princeton University, prepared their own tables to study the question of racial bias. To test the hypothesis that race was responsible for the disparity in sentencing, the defense experts used rigorous statistical techniques to eliminate other variables that could possibly explain the variations attributed to race.
Defense experts updated Table 18 and found that the overall racial disparity first noted in Marshall II has continued at a statistically significant level. That disparity is especially strong at culpability level four. The disparity remained significant, although not as great, at culpability level three, and disappeared at the highest degree of culpability, level five.
Defendant argues that whether one relies on the Special Master’s tabulations or defendant’s own, the effect of impermissible factors — such as race — is most pronounced in cases of “mid-range” culpability, ie., cases in which the death penalty is imposed roughly fifty-percent of the time or less. That finding has considerable support from the academic work of the Special Master, see, e.g., David Baldus et al., Equal Justice and the Death Penalty 14 (1990) (citing work of Baldus demonstrating that “the magnitude of the impact of racial factors on the sentencing outcome varies with the culpability level of the cases”). Those findings also comport with common sense. If invidious racial discrimination enters into a capital-sentencing system, it is likely to do so in the areas in which prosecutorial or juror discretion is greatest. Where the relative culpability of a defendant is either extremely high or extremely low, the effect of prosecutorial or jury discretion is minimized, ie., the likely sentencing outcome is fairly obvious. But in cases of mid-range culpability, in which outcomes vary consistently between life and death, the opportunity for invidious factors to play a role is greater.
Although it does not dispute defendant’s results — including the shocking sixty-five percent disparity — the Court treats defendant’s analysis dismissively. It attacks defendant’s claims primarily *422because of the broad expanse of cases in defendant’s re-worked culpability level four. Ante at 393, 645 A.2d at 714.
The Court is correct in focusing its assessment on the sixty-five percent disparity at culpability level four. Theoretically, all cases grouped at the same culpability level share roughly the same likelihood of receiving a death sentence. The Special Master originally divided the culpability levels into five bands covering ranges of twenty percentage points. Thus culpability level one includes all cases with a 0.00 to .19 predicted frequency, and level two includes .20 to .39 and so on. Because so few death sentences are imposed relative to the death-eligible universe, when the cases are plotted on those levels, the vast majority are grouped at level one. Indeed, under the Special Master’s model only seven cases end up in level four.
The operative hypothesis about race as an invidious factor is that it creeps into the discretionary elements of the system (prosecutorial discretion and juror discretion) in the marginal cases. If racial disparity is going to occur, one would expect it to rear its head not in the obvious cases (extremely high or extremely low predicted frequencies) but in the so-called mid-range cases, between roughly .30 and .70 predicted frequency of death-sentence imposition.
The Special Master recognized that to test for race, one would have to group mid-range cases together. Hence, instead of setting the bands at ranges of twenty percent of expected frequency, the Special Master arranged the entire pool of cases in order from lowest frequency to highest, and then to ensure an adequate number of cases for purposes of assessment, fixed the culpability levels at an even number of cases at each level. At the re-worked culpability level four a wide sweep of predicted frequencies now exists from .14 to .89 among the twenty or so cases grouped there.
Defendant argues that those mid-range cases, omitting the extremely low and extremely high predicted frequencies, are the appropriate test pool for race. The Court reasons, however, that *423by expanding the sweep of level four, defendant is no longer comparing similar cases. Ante at 393, 645 A.2d at 714.
The Court’s position is not without logic as far as it goes. Given such a wide sweep of predicted frequencies, if all the African-American defendants who received a death sentence are at the upper level of that range, i.e., .70 to .80, and all the whites happen to be at the lower ranges, ie., .14 to .30, then relative culpability, not race, explains the disparity. What the Court does not do, however, is examine the range of the eases within culpability level four to see if such a clustering occurs at the upper or lower ends of the scale. That can be done by listing the eases assigned to level four, according to who received death and who life, and including their race, viz:
DEFENDANT: Estimate: Race: Sentence:
1. M. Bey 0^* * * 4 0.19 black death
2. S. Monturi (2) 0.21 white life
3. A. Perry * * * 0.22 black death
4. D. Pitts * * * 0.24 white death
5. Br. Purnell 0.36 black death
6. Jm. Koedatieh 0.40 white life
7. M. Manfredonia 0.42 white life
8. S. Monturi (1) 0.42 white life
9. T. Rose 0.42 white life
10. Wm. Engel 0.43 white/Hisp. life
11. Jos. Guagenti 0.46 white life
12. Geo. Booker (2) 0.47 black life
13. R. Rise (1) 0.48 white life
14. Geo. Booker (1) 0.53 black life
15. Jac. Hightower 0.53 black death
16. M. Melendez 0.54 white/Hisp. life
*42417. S. Moore (1) 0.55 black death
18. S. Moore (21) 0.55 black death
19. Jm. Hunt 0.68 black death
20. L. Reyes 0.72 white/Hisp. life
21. W. Johnson (1) 0.79 white life
22. M. Bey 0.82 black death
23. R. Rise (2) 0.85 white death
Although the range theoretically begins at .14, in fact, the first case is .19. At an anecdotal level, note that three black defendants, Marko Bey, Arthur Perry, and Braynard Purnell, with the first; third; and fifth-lowest predicted frequencies, all received death sentences. Overall, of the twenty-three cases, eight of ten ( Zio) black defendants received the death penalty, but only two of thirteen (2/18) white defendants received death. The most revealing information comes if one accepts the Court’s argument that the range is too big and limits analysis to those cases that fall between .30 and .70, which could be considered a fair mid-range. That sub-grouping has seven black defendants, five of whom received a sentence of death. It also has eight white defendants, all of whom received life sentences. Thus in the most marginal cases, in which one would expect roughly a fifty-fifty chance of receiving a death sentence, one finds that blacks received the death penalty in five out of seven eases and whites in none. Thus, even when the focus of analysis is on the true mid-range cases, those between .30 and .70 predicted frequency of death sentencing, an obvious disparity between races is visible.
Based on defendant’s showing, I am convinced that the racial disparity alleged by defendant at culpability level four is significant and worthy of this Court’s most conscientious and thorough consideration. Yet, the Court exhibits both an uncharacteristic timidity in light of its oft-stated judicial obligation to confront the possibility of invidious racial discrimination and an unwillingness to examine rigorously arguments that credibly and cogently present that possibility.
*425The Court takes the easy route in concluding that too few cases exist from which to make a determination regarding the possible effects of racial discrimination, ante at 388, 645 A.2d at 712. The problems associated with under-sized data pools permeate the whole of the Court’s proportionality review. The inescapable fact remains that all the statistical analyses in this case are disadvantaged by under-sized data pools and consequently large margins for error. That statistical indeterminacy casts a shadow of doubt over the entire proportionality enterprise. The Court, however, determines that defendant alone should bear the burden of dissipating that doubt — a determination that is at odds with the logic of proportionality review itself, which is premised on the notion that defendant has the constitutional right not to be subjected to disproportionate punishment. Moreover, the Court’s placing of the burden of statistical doubt on the defendant contradicts its institutional role in ferreting out race-based discrimination in the administration of justice and its independent responsibility to ensure that invidious discrimination plays no role in determining who receives a death sentence.
Since Ramseur, the Court has warned that it would be vigilant for traces of racial bias in the system. Now confronted by evidence of racial bias, the Court quibbles with methodology and evinces an almost reactionary distrust of statistics. As the Court acknowledges, ante at 386, 645 A.2d at 710, statistical analysis plays an essential role in assessing claims of racial discrimination. See Racial Injustice in the Senate, N.Y. Times May 13, 1994, (Editorial) at A-22 (criticizing members of Senate for refusing to endorse Racial Justice Act appended to House version of federal crime bill, which would expressly authorize statistically-based, racial discrimination studies by federal death-penalty defendants and without which it would not be possible to discover racial discrimination in the administration of the death penalty).
The question is not whether defendant has proved that racial ' discrimination is operative in our capital murder system, but rather, whether defendant’s showing requires the Court to under*426take further investigation and to suspend capital sentencing until allegations of racial discrimination can be thoroughly discounted.
In light of the Court’s wholly inadequate response to the defendant’s statistical arguments, its citation to McCleskey v. Kemp rings with irony. Unless the Court modifies its review of allegations of systematic racial discrimination, it subjects defendants like Marko Bey to the ultimate penalty while waiting for some unspecified quantum of proof that race is acting invidiously in our capital murder system. The Court suffers under the delusion that serious consideration of those allegations can be continually postponed. Time will eventually run out, as it did for Warren McCleskey.
VI
As stated at the beginning, the grave inadequacy of the proportionality review at issue in this case is not the product of a Court lacking in insight or commitment to fairness. Nevertheless, the errors that infect this exercise of proportionality review reflect the fundamental incoherence of our capital murder jurisprudence. The Court’s continued uncritical use of reversed death sentences, an irrationality of the first order, can be supported only by drawing an unfounded distinction between the basic fairness of the sentencing process and the validity of the outcomes of that process. Lacking a coherent and workable methodology, the Court over-relies on a precedent-seeking analysis, which is itself compromised by the harmless-error determinations made in the defendant’s ease on direct review. The Court’s application of frequency analysis is driven by no more than intellectual convenience with little care given to the assessment regarding what is a high or low frequency and lax attention to its functional relationship to precedent-seeking review. Finally, and most disturbingly, evidence of the invidious effect of racial bias in determining who receives death sentences is basically ignored, once again turned aside as premature or inconclusive.
*427Today’s decision serves as further confirmation of the failure of our experiment with capital punishment. The Court’s sincere aspirations for proportionality review cannot be squared with the problems inherent in an attempt to apply an abiding standard of fairness to the imposition of a death sentence. The lights of reason and our common humanity insist that we treat death differently. Sadly, and undeniably, the result of today’s decision is neither rational nor humane. Marko Bey’s now pending execution is bitter testament to the Court’s collective failure to appreciate and respect the limits that inhere naturally in a society governed by law in which the exercise of governmental power is constrained by an insistence that the government act rationally, fairly, and with consistency, or not at all.
The Court’s initial confidence that it could fashion a constitutionally legitimate process for imposing the death penalty, see Ramseur, supra, 106 N.J. at 331, 524 A.2d 188 (“How we will resolve this paradox remains as yet fully unrevealed to us. We shall continue to labor on the process.”) has been shaken by the bewildering experience of proportionality review, the purposes of which, as evidenced by today’s opinion, the Court little understands. I would not hesitate to venture a bet that few in today’s majority would agree with the assurances expressed in Marshall II, that “[o]nce defined ... the process of proportionality review will not ‘frustrate and confuse the Court’.” 130 N.J. at 218, 613 A.2d 1059.
The confusion so readily apparent in today’s decision is the inevitable product of a futile endeavor: the quest to devise and to apply a standard of due process protection commensurate with the gravity of the sentence to be imposed.
We are, by now, inured to the import of the what has become cliched — that “‘death ... is ... different____’” Ramseur, supra, 106 N.J. at 326, 524 A.2d 188 (quoting Lockett v. Ohio, 438 U.S. 586, 605, 98 S.Ct. 2954, 2965, 57 L.Ed.2d 973, 990 (1978)). That phrase is more than a mere truism, however. It serves as the primary evaluative principle in our continuing consideration of the *428feasibility of a death sentence as a constitutionally valid punishment. We noted in Ramseur that proportionality review was to assist the Court in ensuring that “ ‘we have designed procedures [that] are appropriate to the decision between life and death.’” 106 N.J. at 326, 524 A.2d 188 (quoting Pulley, supra, 465 U.S. at 67-68,104 S.Ct. at 888-89, 79 L.Ed.2d at 52). The manifest failure of proportionality review as designed and exercised by this Court is now evident. Equally evident is the impossibility of a capital punishment regime that takes seriously the moral fact that death is indeed different.
The conclusion is clear: the Court must either abandon its mission or accommodate itself to the juridical brutality of imposing death without due process protections commensurate to its awesome finality.
The by-now-familiar argument that the capital murder jurisprudence of the United States Supreme Court rests on two fundamentally incompatible goals was given renewed poignancy by Justice Blackmun’s recent dissent from the Supreme Court’s denial of certiorari in Collins v. Collins, — U.S.-, 114 S.Ct. 1127, 127 L.Ed.2d 435 (1994):
Bather than continue to coddle the Court’s delusion that the desired level of fairness has been achieved and the need for regulation eviscerated, I feel morally and intellectually obligated simply to concede that the death penalty experiment has failed.
[Id. at-, 114 S.Ct. at 1129, 127 L.Ed.2d at 438.]
Justice Blackmun’s conclusion, coming as it does at the close of his long tenure on the Supreme Court, is based on his experiences in attempting, over twenty years, to reconcile the indispensable yet conflicting values of consistency and fairness, a task that he concluded was simply impossible.
The temptation to resolve the conflict by abandoning one value for another is ever-present and apparently irresistible to some. See Walton v. Arizona, 497 U.S. 639, 673, 110 S.Ct. 3047, 3068, 111 L.Ed.2d 511, 541 (1990) (Scalia, J., concurring in part and dissenting in part) (“I will not, in this case, or in the future, vote to uphold an Eighth Amendment claim that the sentencer’s discre*429tion has been unlawfully restricted.”) This Court, however, cannot escape, and should not evade, its obligation to assess the compatibility of our capital-murder regime with the principles of consistency and reliability that undergird our sense of justice when the punishment is death. See Ramseur, supra, 106 N.J. at 185, 524 A.2d 188 (“Sometimes conflicting, the two principles of consistency and reliability reflect the increased demands of accuracy and fairness, rising to constitutional dimension, in the implementation of this unique criminal sanction.”); see also Louis D. Bilionis, Legitimating Death, 91 Mich.L.Rev. 1643, 1684-85 (1993) (arguing that state constitutions are force that can fill void left by Supreme Court’s shortcomings).
As is painfully evident today, our proportionality review is wholly inadequate to aid in legitimating our State’s oft-expressed desire to impose capital punishment. Absent that legitimation, the Court must with Justice Blackmun abandon its quixotic aspiration to reconcile the irreconcilable, or with Justice Scalia draw the conclusion that, “at least one of these judicially announced irreconcilable commands ... must be wrong.” Callins, supra, — U.S. at-, 114 S.Ct. at 1128, 127 L.Ed.2d at 436 (Scalia, J., concurring). Which would the Court renounce: fairness or consistency?
Like Justice Blackmun, retired United States Supreme Court Justice Lewis Powell has also made public his view that constitutional administration of the death penalty is impossible. JOHN C. JEFFRIES, JR., JUSTICE LEWIS F. POWELL, JR., 452 (1994). Justice Powell, who east the deciding vote in favor of executing Warren McCleskey, has since come to conclude that the death penalty cannot be decently administered and that continued attempts to enforce the death penalty will serve only to bring the law into disrepute. Ibid. Today’s decision gives more immediacy to that fear.
As is nearly always true in a capital case, we are impelled to lavish enormous resources, both economic and intellectual, on a person who can lay precious little claim to our sense of compassion. We have, however, an obligation to rely on principle, the *430basic notion that governmental power can be exercised only in fidelity to the values of due process, values that increase as the punitive severity of governmental action escalates. How else can the cornerstones of our collective liberty be secured, unless we are willing to insist that they be sedulously applied on behalf of the least of us?
This Court should acknowledge that no death sentence can be affirmed by a process of review that is beset by inherent contradictions, riddled with subjectivity, laced with error, and tainted by apparent racial bias. If we are uncomfortable in casting Marko Bey as the victim, we can legitimately substitute ourselves. Although perhaps unwittingly, in failing to insist that death be imposed with the full measure of constitutional protection or not at all, we lose a significant and irredeemable part of our civilization built on the rule of law.
APPENDIX
Factual Description of Similar Cases derived from Detailed, Narrative Summary of Death Eligible Case, New Jersey Proportionality Review Project.
1. George Booker:
George Booker, aged thirty-six, went on a “three day crime spree.” After sexually assaulting his first victim and stealing her car, Booker ran down a male pedestrian and stole his wallet. Booker then entered the home of a lesbian couple. He raped and sodomized the first victim, then killed her. Her partner returned home; the defendant forced her to undress and lie in bed next to her partner, then he stabbed her to death. When the victims were found, it was discovered that the first victim’s mouth and forehead had been bashed in, her mouth gagged with a bathrobe tie, and a cord was wrapped around her neck. Police found the defendant carrying a knife in the home of an elderly neighbor. The defendant was convicted of knowing murder with respect to the first victim and purposeful murder of the second. At the *431penalty phase, the jury found the presence of three aggravating factors with respect to the first murder: prior murder, extreme suffering, and contemporaneous felony. The jury found three aggravating factors for the second victim, prior murder, extreme suffering, and murder to escape detection. The jury also found that the mitigating factors of emotional disturbance and the catchall factors applied to both murders. The jury deadlocked on whether the mitigating factors outweighed the aggravating factors, and Booker was sentenced to an aggregate sentence of life imprisonment with a sixty year parole disqualifier.5
2. Carlos Vasquez:
Carlos Vasquez, aged forty-three, abducted a thirteen year old, raped her and killed her. The defendant had bound her hands and feet together and pulled them behind her back. The cause of the victim’s death was asphyxia by gagging, ligature strangulation, and fracture of the cervical spine. The defendant denied mental health or substance abuse problems. The jury found the catch-all mitigating factor. Defendant received an aggregate sentence of life imprisonment with a forty year parole disqualifier.6
3. Leroy Taylor:
LeRoy Taylor, age twenty-five, had raped and strangled a thirteen-year-old babysitter hired by his girlfriend. Taylor had a prior conviction for the murder of a four-year-old girl. He pled *432guilty to the murder and did not proceed to a penalty phase, apparently as part of the plea agreement. Taylor received a life sentence with a total parole disqualifier of thirty-two-and-one-half years.
4. James Koedatich:
James Koedatich, age thirty-four, abducted an eighteen-year-old Amie Hoffman from the parking lot of the Morris County Mall at approximately 9:30 p.m. on November 23, 1992. The defendant raped and sodomized the girl, stabbed her multiple times, including two fatal chest wounds, then left her body in an isolated holding tank in Randolph Township. Defendant had committed a prior murder in Florida ten years earlier, and another murder of a twenty-five-year-old woman two weeks after the Amie Hoffman murder. At a retrial of the penalty phase, jurors found four aggravating factors: 4(a), prior murder; 4(c), extreme suffering; 4(g), contemporaneous kidnap and sexual assault, and 4(f), murder to escape detection. The jury found a factor of childhood trauma under the 5(h) catch-all factor. During the weighing process, the jury deadlocked and life imprisonment was imposed.7
5. Anthony McDougald:
Anthony McDougald committed multiple murders involving burglary and sexual assault of at least one of the victims. Defendant, age twenty-seven, had been dating a thirteen-year-old girl, the daughter of the two victims. Her parents objected to Antoinette and McDougald having sexual relations. McDougald entered their home with another thirteen-year-old girl, slashed Mr. Bass’ *433throat and stabbed him in the chest, and struck him in the head with a baseball bat. He then struck the mother with a cinder-block and a bat, cut her throat, and inserted the bat into her vagina. The thirteen-year old girl also participated in the killings. At the penalty trial, the defense sought to portray the defendant as a product of a violent and deprived childhood, and severely despondent over the impending divorce from his wife. MeDougald’s mother testified at trial that when the defendant was a child, her sister, who had been living with them in North Carolina, had physically abused the defendant. A few years later, when the family moved to Newark, New Jersey, defendant was also beaten by his mother’s boyfriend and repeatedly witnessed his mother being beaten. At the penalty trial, the jury found aggravating factors 4(e), intent to cause suffering; 4(f), murder to escape detection; and 4(g), commission of the murders while engaged in commission of a burglary. The jury also found mitigating factors 5(a), influence of extreme mental or emotional disturbance, and 5(h), the catch-all factor concerning background and character. The jury found that aggravating factors outweighed mitigating factors. Retrial of the penalty phase, based on the court’s erroneous charge on the 4(c) aggravating factor, is now pending.8
For affirmance — Chief Justice WILENTZ, and Justices CLIFFORD, POLLOCK, O’HERN, GARIBALDI and STEIN— 6.
For reversal — Justice HANDLER — 1.
Prior to the Court’s decision in Marshall II, the Legislature amended the Capital Punishment Act to provide that only "similar cases in which a sentence of death has been imposed” would form the basis of comparison for the purpose of proportionality review. L. 1992, c. 5 (effective May 12, 1992) (codified at NJ.S.A. 2C:ll-3(e)). The Court does not apply the amendment to defendant's case, nor does it consider its constitutionality. Ante at 344, 645 A.2d at 690.
Limiting proportionality review to death-sentenced cases is irrational and destroys the analytic value of proportionality review itself. Review of prosecutorial discretion would have to be effectively abandoned. Detecting invidious discrimination, like racial bias, within the amended statute's narrow universe, will be almost impossible. The former Chief Justice of the Nebraska Supreme Court recognized that fact, pointing out that limiting proportionality review to death-sentenced cases is like reviewing alleged discrimination in public transportation by looking only at those riding in the back of the bus. See State v. Palmer, 224 Neb. 282, 399 N.W.2d 706, 752 (1986) (Krivosha, C.J., concurring and dissenting).
The State, however, in its briefs, was willing to speculate on why this jury did not find age as a mitigating factor. It suggested that perhaps the jury was swayed by the thought that if given a life sentence carrying a mandatory thirty-year parole ineligibility period, Bey, being relatively young, aged eighteen, might be released at the age of forty-eight. The jury, had it so reasoned, would have been wrong; Bey, given his prior conviction and sentence would not have been eligible for release during his life time. The jury was not accurately informed of the true implications of a life sentence for Bey, however. The sentencing court failed to answer its inquiry on the matter, a failure that this Court held to be harmless error. Bey, supra, 129 N.J. at 606, 610 A.2d 814.
Nor does the distinction between "structural” and "trial” error measurably aid the accuracy or enhance the fairness of harmless-error review in the penalty phase of capital sentencing. See Charles J. Ogletree, Jr., The Harm of Applying Harmless Error to Coerced Confessions, 105 Harv.L.Rev. 152, 159-64 (1991).
* * * indicates cases in which the Court later determined that the defendant was not even death eligible, e.g., Bey I in which the Court determined that he was a juvenile. Those names do not appear in any of the finished tables used by defendant or the Special Master, but for purposes of race analysis, those cases are included making a total of 23 cases in level 4. The majority criticizes the use of those three cases. Ante at 393, 645 A.2d at 714. Yet, of course, the majority, heedless of the irony, steadfastly insists that reversed death sentences, like those in Bey I, continue to represent "a societal consensus concerning the deathworthiness of a defendant.” Ante at 348, 645 A.2d at 692.
Defendant argues that Booker was clearly more culpable because Booker had killed a total of three persons. Booker’s level of victimization was higher than Bey’s, since the 4(c) factor of aggravated assault or torture was found to be present. Unlike Bey, Booker, though he had come from a large sharecropping family, had not suffered an abused childhood. In addition, Booker was considerably older than Bey.
Defendant argues that the levels of victimization in the two cases are equivalent but that Vasquez's older age, lack of mental disturbance, and lack of known abusive childhood, and the youth of the victim clearly make him more culpable.
Defendant maintains that the fact that Koedatich had committed a third murder ten years earlier, and was 34 years old when he committed the New Jersey murders, renders him more culpable than Marko Bey. The defense contends that Koedatich received a life sentence because he received a fair trial, in as much as, that the jury received "a full picture of his abusive childhood,” understood the alternatives to the death sentence, and was shielded from irrelevant and inflammatory evidence.
The defense claims that McDougald did not experience the extended abuse as a child that was inflicted on Marko Bey, and that Marko Bey did not engage in such extensive torture or mutilation of his victim.