Frames:
JUDGING THE EFFECTIVENESS OF THE SUPREME COURT'S DEATH PENALTY JURISPRUDENCE ACCORDING TO THE COURT'S OWN GOALS: MILD SUCCESS OR MAJOR DISASTER?

DAVID MCCORD[*]
Copyright 1997 Florida State University Law Review

I. INTRODUCTION
II. STEIKER AND STEIKER'S MISTAKE REGARDING UNDERINCLUSION
A. Steiker and Steiker's Argument
B. Underinclusion Depicted
C. The Court's Pronouncements on Which Academic Underinclusionists Rely
D. The Causes of Underinclusion
E. The Legitimacy of Merits-Based Underinclusion
1. Merits-Based Underinclusion at the Guilt Phase
2. Transferring the Underinclusion Urge to the Death Sentencing Stage
3. Merits-Based Underinclusion Through State Death-Sentencing Procedures
F. Supreme Court Underinclusiveness Jurisprudence
1. The Court Has Never Entered the Fray with Respect to Prosecutorial Underinclusion Due to Mundane Causes.
2. The Court Has Prohibited States from Curtailing Guilt-Determiners' Power to Exercise Merits-Based Underinclusion
3. The Court Has Prohibited States from Undermining the Possibility of Merits-Based Sentencer Underinclusion (and, Indirectly, Has Fostered Merits-Based Prosecutorial Underinclusion)
4. The Court's Real and Only Underinclusion Concern: Invidious Underinclusion
G. The Court's Primary Concern: Minimizing Overinclusion
III. STEIKER AND STEIKER HAVE NOT PROVEN THAT THE COURT'S REGULATORY EFFORT HAS BEEN AN INEFFECTIVE DISASTER AND HAVE IGNORED EVIDENCE TO THE CONTRARY
A. The Flaw in Steiker and Steiker's Premise
1. The Winnowing Effect of Statutory Narrowing of Death-Eligibility
2. The Effect of Nonperfunctory Sentencing Hearings
3. Has the Court Succeeded in Minimizing Overinclusion Due to Race?
B. The Flaw in Steiker and Steiker's Conclusion
1. Empirical Evidence of the Decline in Overinclusion
a. BWP's Georgia Study
b. Some New Evidence That Overinclusion Is Not a Widespread Defect in Georgia's System
2. Empirical Evidence of the Decline in Underinclusion
C. Are the Post- Furman Declines in Overinclusion and Underinclusion Attributable to the Court's Jurisprudence?
IV. CONCLUSION

I. INTRODUCTION

Opponents of the death penalty have long taken the United States Supreme Court to task for not ruling that the penalty is per se unconstitutional. But there also has been a longstanding breed of less absolutist critics. These critics are willing to assume arguendo that regulation rather than abolition is a proper stance for the Court. They then argue that the Court's chosen means of regulation have proven ineffective to remedy the very evils that prompted the Court to undertake regulation of capital punishment in the first place. The most complete and high-profile presentation of this critique occurs in a provocative article in a recent issue of the Harvard Law Review.[1] There, Professors Carol S. and Jordan M. Steiker seek to perform an "internal" critique[2] of Supreme Court death penalty jurisprudence: an examination of whether the Court has achieved the goals that it has set for itself since it began to regulate state death penalty law more than two decades ago.[3] After exhaustive and thoughtful analysis, the authors come to a damning conclusion: "[T]he Supreme Court's chosen path of constitutional regulation of the death penalty has been a disaster, an enormous regulatory effort with almost no rationalizing effect."[4] If Steiker and Steiker are correct, they have succeeded in exposing the fact that the Court's regulatory effort is the juridical equivalent of the Emperor's new clothes-costly, yet embarrassingly ineffective for the goals Steiker and Steiker attribute to the Court: decreasing "overinclusion"-the imposition of death sentences on defendants who are not among the "worst" murderers; decreasing "underinclusion"-the imposition of death sentences on only some, rather than all, equally culpable murderers; and increasing "individualization" of sentencing by considering all aspects of the defendant's character; all through "heightened procedural reliability."[5]

My purpose is to dispute the arguments of the less absolutist critics represented by Steiker and Steiker. I will do this by responding to Steiker and Steiker's piece, which I deem to be the "state of the art" for this type of critique. I will argue that despite the virtues of their article,[6] Steiker and Steiker are wrong in two important ways.

First, they, along with other critics I will call "academic underinclusionists," have wrongly concluded that minimizing all underinclusion has been one of the Court's concerns. This incorrect conclusion regarding underinclusion is exacerbated by Steiker and Steiker's view that the Court's three alleged substantive concerns-overinclusion, underinclusion, and individualization-are of equal weight.[7] I will argue, rather, that there are three causes of underinclusion: (1) "merits-based"-the case is underincluded because reasonable minds can differ with respect to whether it is one of the "worst" homicides; (2) "mundane"-the case is underincluded because of everyday concerns that apply to capital and noncapital cases alike, such as loss of a key witness, case overload leading to plea bargaining, etc.; and (3) "invidious"-the case is underincluded because of illegitimate racial or class bias. I believe that the Court[8] has viewed merits-based underinclusion as a desirable phenomenon: it has often prohibited states from using mechanisms that would curtail such underinclusion. As to mundane underinclusion, I believe the Court has recognized that it is beyond the Court's power to regulate, short of complete abolition of the death penalty, a step the Court has never been willing to take. That leaves invidious underinclusion, the only form of underinclusion about which the Court has been concerned. The academic underinclusionists' mistake regarding the Court's alleged across-the-board concern with underinclusion is a serious one because underinclusion is a far more prevalent feature of capital punishment than is overinclusion. Thus, if these critics are wrong about the Court's concern with underinclusion, it throws their entire argument into serious doubt.

The correct view, I will contend, is that the Court has had only one primary goal for its regulation of capital punishment: decreasing overinclusion, with particular interest in minimizing invidious overinclusion due to racial bias. I will argue that the Court's regulatory efforts have prompted responses from the states that, at least on their faces, seem to have the potential to partially achieve the Court's goal.

The question then becomes, of course, whether these state responses have actually succeeded in producing such improved results. I will contend that the second way in which I believe Steiker and Steiker have erred is by not substantiating their claim that the Court has legitimated state death penalty systems having the same potential to operate arbitrarily as the pre-Furman v. Georgia[9] systems. Beyond that, I will argue that they have not proven that post- Furman systems are in fact systematically operating in as arbitrary a fashion, even accepting Steiker and Steiker's assertion that the Court is opposed to all underinclusion. In fact, the best available evidence strongly suggests that post- Furman systems are operating less arbitrarily with respect to both overinclusion and underinclusion,[10] and that this improvement is likely due in significant part to the Court's regulatory efforts.[11]

Before I begin my task, though, a bit of a disclaimer is in order. The analysis that Steiker and Steiker perform, and that I will undertake in response, is of a very traditional scholarly sort-trying to figure out what a body of legal doctrine says, what the policy goals behind it are, and whether it seems to be fulfilling those goals. The same type of analysis could be applied to a body of contract, tort, or property law. But the death penalty seems different to many people: it evokes such intense emotions in some opponents that scholarly analysis seems beside the point. Lest those opponents take me as an apologist for the death penalty, let me state my position for the record: I do not believe the death penalty should exist. Nor do I want to be understood as arguing that the Court has done an excellent job in choosing and implementing the goals it is seeking to foster through its death penalty jurisprudence. I do insist, though, that the topic is fair game for traditional, dispassionate legal analysis, that such analysis in this instance turns out to be quite illuminating, and that when the analysis cuts in the Court's favor, one must call it as one sees it.

II. STEIKER AND STEIKER'S MISTAKE REGARDING UNDERINCLUSION

A. Steiker and Steiker's Argument

Steiker and Steiker explain their mission as follows:

We thus have chosen to read Furman and Gregg [v. Georgia, Proffitt v. Florida, Jurek v. Texas, Woodson v. North Carolina, and Roberts v. Louisiana] together as a way of identifying the Supreme Court's concerns and goals regarding its constitutional regulation of capital punishment. We stop with Gregg and its companion cases, however, because the seeds of all of the rest of the Court's capital jurisprudence can be traced to the themes that it sounded in 1972 and 1976. . . . [I]t seems likely that the Court itself would think it fair to measure the success of its capital punishment jurisprudence against the concerns articulated in Furman and Gregg et al. Given that our critique of the Court is an internal one, identifying these concerns becomes a crucial part of our project.[12]
As the authors themselves recognize, it is crucial to their enterprise to correctly identify the concerns that the Court would itself choose to analyze its performance. Here, then, are the concerns Steiker and Steiker identify: "We think that these concerns can fairly be grouped around four ideas: desert (the problem of overinclusion), fairness (the problem of underinclusion), individualization, and heightened procedural reliability."[13] The authors describe each of these concerns more fully. Overinclusion is "the application of the death penalty in circumstances in which, notwithstanding the sentencer's decision, the sentence is not deserved according to wider community standards,"[14] because the defendant is not among the "worst" offenders.[15] Underinclusion is the "failure to treat equally deserving cases alike,"[16] which is unacceptably unfair when the system "treat[s] others, just as 'deserving' as the condemned defendant, more leniently for no reason, or for invidious reasons."[17] Individualization refers to the idea that determination of the "worst" crimes requires inquiry not only into facts regarding the offense, but also facts regarding the offender that might be cause for judging the offender to not be among the "worst."[18] Heightened procedural reliability is the means to these other three ends: "[R]eliability is tied to strong notions of desert and fairness among defendants. Hence, one can view the Court's commitment to heightened procedural reliability as its manner of making good on its three substantive commitments-to desert, fairness, and individualization in capital sentencing."[19]

Having identified these concerns, Steiker and Steiker go on to demonstrate how, in their view, the Court's jurisprudence has failed to result in improved capital punishment systems that assuage any of the four concerns. As to overinclusion, Steiker and Steiker claim that the Court has failed to require states to narrow significantly the class of death-eligible defendants,[20] and has not itself placed any important limitations on death-eligibility[21] (other than the ban on the death penalty for rape in Coker v. Georgia[22]). As to underinclusion, the authors argue that the Court has failed to require the sentencer's decision to be "channeled"[23] at the "critical moment when it decide[s] whether to impose a death sentence."[24] This leaves standardless discretion in the sentencer, which inevitably leads to like cases being treated differently.[25] The authors also are critical of the Court's approach to achieving its goal of individualization.[26] Although the Court has certainly opened the door to all individualizing mitigating evidence, it also has concomitantly permitted arbitrariness into the process by the very same means.[27] Nor has the Court required states to channel the sentencer's consideration of individualizing evidence.[28] Finally, as to heightened procedural reliability, the Court has failed, according to Steiker and Steiker, to go as far as it should in requiring states to treat death differently procedurally. The Court has imposed some special protections,[29] but "it has done so in an entirely ad hoc fashion and left untouched a substantial body of doctrine that relegates capital defendants to the same level of protection as noncapital defendants."[30] Thus, Steiker and Steiker contend that the Court's capital jurisprudence has been an ineffective disaster.[31] However, I believe Steiker and Steiker have erred in their identification of the Court's alleged concern with underinclusion, which substantially undermines the subsequent critique.

B. Underinclusion Depicted

Steiker and Steiker are certainly operating on the basis of received wisdom when they contend that the Court is concerned with the problem of underinclusion. Several other academic underinclusionists have concluded that the Court views underinclusion as a vice.[32] The most well-known are Professors David Baldus, George Woodworth, and Charles Pulaski (hereinafter "BWP"), who refer to the same problem as that of "excessiveness,"[33] or, even more descriptively, "comparative excessiveness."[34] BWP performed thorough and influential research on a body of several hundred pre- and post-Furman cases to study the extent of the underinclusion problem in Georgia.[35] It is crucial to understand what academic underinclusionists mean when they refer to underinclusion, particularly as it differs from overinclusion. While I will discuss the BWP statistical work later,[36] the easiest visual aid to use in understanding the issue is one created by Professor Arnold Barnett about a decade ago.[37]

Professor Barnett examined Georgia murder convictions[38] between March 28, 1973, and June 30, 1978,[39] and analyzed the variables that appeared to him to account for whether or not death sentences were imposed.[40] He concluded that the likelihood of a death sentence depended upon how the case scored on each of three variables: certainty that the killing was deliberate; relationship between defendant and victim; and vileness of killing (which involved two sub-inquiries: whether there was a plausible claim of self-defense, and whether the killing was particularly horrific in any of fourteen specified ways).[41] He developed a protocol for each of the three variables that permitted him to assign a score of 0, 1, or 2 as to the first and third variables, and 0 or 1 as to the second (0 having little or no tendency to lead to a death sentence, 1 having a moderate tendency, and 2 having a strong tendency-thus, the best (least death-prone) score a case could get was 0, 0, 0, and the worst (most death-prone) score a case could get was 2, 1, 2).[42] He then analyzed what proportion of defendants in cases having identical scores received the death sentence.[43] His chart is reproduced in Figure 1,[44] although I have added letters to each box for easy reference. The score of each case on Barnett's three variables is in parentheses on the bottom line of each box. The first number on the top line is the proportion of cases in that box in which the defendant was sentenced to death. Next to the proportion, in parentheses, is the fraction on which the proportion is based-the numerator is the number of defendants in that category who received the death sentence, and the denominator is the number of cases in that category. For example, in box K, there were fifty-nine defendants whose cases scored 1, 1, 1, and fifteen of them received the death sentence, a proportion of .25.

FIGURE 1



B
.0
(0/1)
(0,1,0)

E
.05
(1/22)
(0,0,2)

J
.09
(1/11)
(2,0,1)





A
.0
(0/3)
(0,0,0)

C
.03
(1/31)
(1,0,0)

F
.0
(0/9)
(1,1,0)

K
.25
(15/59)
(1,1,1)

N
.81
(59/73)
(1,1,2)

P
.86 (6/7)
(2,1,2)



D
.0
(0/42)
(0,0,1)

G
.02
(1/41)
(0,1,1)

L
.30
(6/20)
(0,1,2)

O
.56
(10/18)
(2,0,2)







H
.01
(1/106)
(1,0,1)

M
.28
(11/39)
(1,0,2)









I
.0
(0/1)
(2,0,0)












There are three important things to note about Barnett's diagram. First, it generally reflects a common-sense continuum, from a nonexistent or very low death sentencing rate in the lower-scored cases (boxes A-J), through a significant but not preponderant rate in moderate-scored cases (boxes K-M), to a preponderant rate in high-scored cases (boxes N-P). The second thing to note is how well the diagram illustrates the clear cases of overinclusion: unless Barnett's analysis failed to pick up on some key aggravating aspect of the cases-always a possibility when using a factors-type analysis-the imposition of the death sentence on one defendant in each of boxes C, E, G, H, and J seems clearly to be arbitrary in that it is unlikely that those defendants were among the worst murderers in Georgia during that time span. For those defendants, Justice Stewart's famous aphorism that "[t]hese death sentences are cruel and unusual in the same way that being struck by lightning is cruel and unusual"[45] seems apt.

The third important thing to note about Barnett's diagram is how prevalent underinclusion is compared with overinclusion. I draw the dividing line separating overinclusion and underinclusion between boxes J and K. In all the boxes A-J, the death sentencing rate is below ten percent, while in all the boxes K-P, the death sentencing rate is at least twenty-five percent. To me, this indicates that the death-sentenced defendants in boxes K-P were not overincluded-their crimes were of a nature that evoked death sentences with regularity. If my dividing line is right, then there are 109 instances of underinclusion in boxes K-P, compared with only five cases of overinclusion in boxes A-J. This means that underinclusion is a much more prevalent aspect of the system than is overinclusion.

Academic underinclusionists do not demand perfect evenhandedness of the system. BWP set the threshold for evenhanded sentencing as above .80, i.e., in over eighty percent of similar cases, the sentence was death.[46] Thus, they are willing to accept the evenhandedness of the death sentences in cases in boxes P (.86 ratio) and N (.81 ratio). The true underinclusion problem, according to academic underinclusionists, inheres in boxes K-M and O, where there is both a significant death sentencing rate and a significant number of defendants who are spared the ultimate sanction.[47] The diagram thus well illustrates the core concern of the academic underinclusionists (including Steiker and Steiker)-a concern that they believe the Court shares. And, admittedly, academic underinclusionists do have significant language from the Court's opinions on which to rely.

C. The Court's Pronouncements on Which Academic Underinclusionists Rely

Academic underinclusionists rely upon a series of Supreme Court pronouncements as the basis for their argument that the Court has been concerned with underinclusion.[48] These pronouncements begin in three of the concurring opinions in Furman. In one, Justice Douglas opined:

The high service rendered by the "cruel and unusual" punishment clause of the Eighth Amendment is to require legislatures to write penal laws that are evenhanded, nonselective, and nonarbitrary, and to require judges to see to it that general laws are not applied sparsely, selectively, and spottily to unpopular groups.[49]
Justice Stewart stated, "For, of all the people convicted of rapes and murders in 1967 and 1968, many just as reprehensible as these, the petitioners are among a capriciously selected random handful upon whom the sentence of death has in fact been imposed."[50] And Justice White argued:

But when imposition of the penalty reaches a certain degree of infrequency, it would be very doubtful that any existing general need for retribution would be measurably satisfied. Nor could it be said with confidence that society's need for a specific deterrence justifies death for so few when for so many in like circumstances life imprisonment or shorter prison terms are judged sufficient . . . .[51]
The Court reiterated these sentiments in upholding new death penalty regimes four years later. In Jurek v. Texas,[52] the Court noted that "Texas has provided a means to promote the evenhanded, rational, and consistent imposition of death sentences under law."[53] In Proffitt v. Florida,[54] the Court approved judge rather than jury sentencing by noting:

[I]t would appear that judicial sentencing should lead, if anything, to even greater consistency in the imposition at the trial court level of capital punishment, since a trial judge is more experienced in sentencing than a jury, and therefore is better able to impose sentences similar to those imposed in analogous cases.[55]
And, in Gregg v. Georgia,[56] the Court cited favorably to the White and Stewart opinions in Furman.[57]

Similar sentiments appear in later opinions. In Godfrey v. Georgia,[58] the Court found the aggravating circumstance of an "outrageously or wantonly vile, horrible or inhuman" offense unconstitutionally vague, stating, "There is no principled way to distinguish this case, in which the death penalty was imposed, from the many cases in which it was not."[59] In Gardner v. Florida,[60] the Court stated the general principle that "the State must administer its capital-sentencing procedures with an even hand . . . ."[61] And in Zant v. Stephens,[62] the Court noted that the findings that the defendant had escaped from custody and had a prior conviction for a capital felony "adequately differentiate this case in an objective, evenhanded, and substantively rational way from the many Georgia murder cases in which the death penalty may not be imposed."[63] Reading these statements, one would have to agree with the academic underinclusionists' conclusion that the Court is concerned with underinclusion. Nevertheless, to really understand what the Court is getting at by these statements, one must recognize that there are three different reasons for underinclusion, and that the Court views the validity of these reasons in quite different ways.

D. The Causes of Underinclusion

Steiker and Steiker suggest that there are only two possible reasons for the "failure to treat equally deserving cases alike":[64] "for no reason, or for invidious reasons."[65] Obviously, both of these are illegitimate reasons-no rational person wants to see a defendant sentenced to death because of the result of a coin toss, or because of racial or class bias. However, the landscape of underinclusion has more features than these two, as will be apparent by examining a set of Georgia cases that have been studied extensively.

During the five-year period covered by Barnett's study, there were about 2,000 defendants in Georgia who were convicted of murder or voluntary manslaughter.[66] Of these, 594 defendants received either life or death sentences after jury trials, or were sentenced to death after pleading guilty to murder.[67] These 594 formed the basis for Barnett's analysis.[68] In 194 of those cases, there was a single death penalty trial, and in twelve additional cases, there were two or more penalty trials.[69] The upshot is a total of 206 penalty trials with 112 death sentences imposed.[70] The obvious challenge arising from the data is to explain the factors that resulted in some defendants being sentenced to death, and others not. Barnett and BWP made herculean efforts to isolate factors about the cases from which explanations could be inferred. I will now, instead, identify factors about the system that account for the underinclusion.

Only about thirty percent of homicide cases (594 out of approximately 2,000) resulted in murder convictions.[71] Thus, there was a seventy percent underinclusion rate "off-the-top" as to whether the death sentence would be sought. This underinclusion must be due in part to what I will call "prosecutorial charging underinclusion"-the prosecutor chose not to seek a murder conviction. The reasons for prosecutorial underinclusion can conceptually be divided into three categories: (1) merits-based-the prosecutor did not personally believe that the case warranted a murder conviction, and/or the prosecutor did not believe the jury could be convinced to return a murder verdict; (2) mundane-which includes a whole host of variables that are endemic to a prosecutor's job, e.g., a key witness disappeared or changed stories, the prosecutor had a case overload and needed to plea bargain, the prosecutor had to offer a deal to one co-defendant to obtain testimony against another, etc.; and (3) invidious-the prosecutor's decision not to seek a murder conviction was based on illegitimate discriminatory motives, most likely racial or class bias.

The seventy percent off-the-top underinclusion rate also must be in part a result of what I will call "guilt-determiner underinclusion"-the prosecutor sought a murder conviction, but the trier of fact (usually a jury) returned a manslaughter verdict. This underinclusion can be divided into two categories: (1) merits-based-the trier of fact felt that the facts warranted the lesser conviction; and (2) invidious-the trier of fact reduced the level of the conviction for illegitimate discriminatory motives.

Of the 594 cases that resulted in murder convictions, prosecutors sought death sentences in only about one-third of them.[72] Of the 400 instances in which the death penalty was not sought, 123 can be explained by the fact that the defendant was not death-eligible because there was no statutory aggravating factor present.[73] This leaves about 270 cases of what I will call "prosecutorial sentencing underinclusion." Again, the universe of possible causes of prosecutorial underinclusion can be divided into three categories: merits-based, mundane, and invidious. Merits-based underinclusion in sentencing means that the prosecutor personally did not believe the case warranted a death sentence, and/or did not believe the sentencer would return a death sentence. "Mundane" in this context refers to non-merits-based, noninvidious causes, such as that the county could not afford to prosecute a death penalty case,[74] the victim's family was opposed to imposition of a death sentence, the prosecutor agreed to give a convicted co-defendant a deal to obtain penalty phase testimony against another co-defendant, etc. "Invidious" refers again to decisions based on illegitimate motives, such as racial and class bias.

The data also show that the sentencers (presumably usually a jury, although a defendant can waive jury sentencing and opt to be sentenced by the judge)[75] declined to impose death sentences in ninety instances out of 203, a "sentencer underinclusion" rate of about forty-five percent. This underinclusion can theoretically be divided into the same two categories as guilt-determiner underinclusion: merits-based-the sentencer felt that the case did not warrant a death sentence, and invidious-the sentencer declined to impose a death sentence for illegitimate motives. The diagram in Figure 2 may help to summarize the flow of underinclusion. This diagram describes a richer and more true-to-life understanding of underinclusion than is presented by Steiker and Steiker. The richer understanding of underinclusion has important implications for understanding the Court's death penalty jurisprudence. A necessary prelude to examining that jurisprudence, though, is a paean to the virtues of merit-based underinclusion.

E. The Legitimacy of Merits-Based Underinclusion

1. Merits-Based Underinclusion at the Guilt Phase

Our system is designed to be merits-based underinclusive at the guilt/innocence stage. The old adage that it is better for ten guilty persons to go free than for one innocent person to be convicted[76] is premised on the belief that being convicted of a crime is qualitatively different and more drastic than any other action the government can take against an individual.[77] When the individual's stake is so high, the risk of error should fall on the side of underinclusion by a fairly large margin.[78] This underinclusion sentiment is reflected in the criminal justice system by two primary devices, which I will refer to as "traditional underinclusion mechanisms": the requirement of proof of guilt beyond a reasonable doubt, and the requirement of a supermajority vote of the jury for conviction (unanimity in most jurisdictions). The Court has held that both of these traditional underinclusion mechanisms are mandated by the Constitution,[79] which means they must reflect a laudable policy goal. Thus, at the guilt/innocence stage, merits-based underinclusion is indisputably a virtue, not a vice.

2. Transferring the Underinclusion Urge to the Death Sentencing Stage

The virtue of merits-based underinclusion at the sentencing stage in general is a less straightforward proposition. Where the system is, by definition, not dealing with innocent persons, there are undeniable benefits to consistency in terms of actual fairness, the appearance of fairness, and general deterrence. Equally undeniably, however, there should always be a place for individualization in the sentencing process, as even quite rigid sentencing guidelines recognize.[80] The clearest underinclusion mechanism at sentencing is simple mercy[81]-the defendant may deserve a certain sentence, but there is nonetheless some reason, perhaps inarticulable, not to give it to the defendant.

As to capital sentencing, the urge for merits-based underinclusion is strong, and its wellsprings parallel those that animate the desire for underinclusion at the guilt/innocence stage. The state's extinguishing of the defendant's life seems qualitatively different and more drastic than any other punishment that can be imposed-to most of us, death is different.[82] When the individual's stake is so high, the risk of error should fall on the side of underinclusion. Thus, it would not be surprising to find that many states have chosen to be merits-based underinclusive as to capital sentencing. Further, it would be quite surprising to find that the Court has been antipathetic to such underinclusion. Let us examine the statutes, and then the case law.

3. Merits-Based Underinclusion Through State Death-Sentencing Procedures

State death sentencing procedures are chock-full of underinclusion mechanisms. These devices are at their most potent in states where the jury is the sentencer. Twenty-eight of the thirty-eight death penalty states fall into this category.[83] In eighteen of those states, the jury must unanimously find that at least one statutory aggravating circumstance has been proven by the state.[84] If there is unanimous agreement, the jury then proceeds to consider the aggravating circumstances in comparison to the mitigating circumstances to determine sentence. At this stage, in all twenty-eight jury sentencing states, the jury must unanimously vote for a death sentence.[85] It is important to note that in all but two of the twenty-eight states,[86] a failure to achieve unanimity at either point results in the sentence defaulting to life imprisonment.[87] Thus, unlike a hung jury at the guilt/innocence phase, which permits a retrial at which the defendant is again at risk of conviction, in twenty-six states, a hung jury at the death sentencing proceeding does not permit another sentencing hearing before another jury where the defendant is again at risk of receiving a death sentence.

The beyond-a-reasonable-doubt standard is almost, but not quite, as prevalent as the unanimity requirement. Twenty-five states require that at least one aggravating circumstance be proven beyond a reasonable doubt.[88] Six states also require that the aggravating circumstances so found outweigh the mitigating circumstances beyond a reasonable doubt.[89] Some states that do not use the beyond-a-reasonable-doubt standard in the determination of sentence nonetheless use a more moderate underinclusion mechanism by placing the burden on the state (presumably by a preponderance standard) to prove that the aggravating circumstances outweigh the mitigating.[90]

Most states employ these traditional underinclusion mechanisms in combination. The most underinclusive combination, found in three states, is a unanimity requirement as to the aggravating circumstances, coupled with a requirement that the aggravating circumstances unanimously be found to outweigh the mitigating circumstances beyond a reasonable doubt.[91] The most common combination, found in eleven states, is a unanimity requirement as to the aggravating circumstances that must be found beyond a reasonable doubt, followed by a requirement that the jury unanimously find for death after considering the aggravating and mitigating circumstances.[92] Other combinations exist.[93] The upshot is that every jury-sentencing state has opted to employ at least one, and usually more than one, underinclusion mechanism.

One of the ten non-jury-sentencing states, Nevada, has a unique system, but one that makes substantial use of the traditional underinclusion mechanisms. There, if the jury cannot unanimously find at least one aggravating circumstance, or cannot unanimously agree to a death sentence, the decision passes to a three-judge panel, which can only impose a death sentence if it unanimously finds at least one aggravating circumstance and unanimously agrees to impose a death sentence.[94] Otherwise, the sentence defaults to one of imprisonment.[95] Recently, Colorado changed its law to require sentencing by a three-judge panel with no jury involvement; however, the three judges must find an aggravating factor beyond a reasonable doubt and must be unanimous in deciding for death-absent unanimity, the sentence defaults to imprisonment.[96]

The remaining eight non-jury-sentencing states give the sentencing decision to the trial judge.[97] The unanimity requirement is, of course, inapplicable to a one-person sentencer. However, some of these states do employ burden of persuasion underinclusion mechanisms. Two states require the judge to find the existence of the aggravating circumstances beyond a reasonable doubt.[98] In one of these states, the judge must find that the aggravating circumstances outweigh the mitigating circumstances beyond a reasonable doubt.[99] Four states employ a moderate underinclusion mechanism on the sentence determination question, requiring that the judge must find that the aggravating circumstances outweigh the mitigating circumstances (presumably by a preponderance standard).[100] Four states, though, do not employ any underinclusion mechanism in judge sentencing.[101]

There is no definitive research examining whether death sentencing underinclusion mechanisms have an actual underinclusive effect, and, if so, how much.[102] Yet, common sense suggests that there is such an effect. In particular, it seems that the requirement for unanimity in jury sentencing states should have a powerful underinclusive effect. There is, in fact, some data to support this hunch: death sentencing rates tend to be lower in jury sentencing states than in judge-sentencing states.[103]

Georgia, a jury-sentencing state, employs the most common combination of traditional underinclusion mechanisms. To impose a death sentence, the jury must unanimously find beyond a reasonable doubt that at least one aggravating circumstance exists, and then, after considering the aggravating and mitigating evidence, must unanimously decide that death is the appropriate sentence.[104] Since there are three causes of underinclusion-merits-based, invidious, and mundane-Barnett's chart[105] becomes at once both more problematic and more meaningful. Now, rather than simply labeling the disparities in boxes K-M and O, where the death sentencing rates range from .25 to .56, as underinclusive and thus undesirable, we have to ask what kind(s) of underinclusion is/are at work. If, as I have argued, underinclusion mechanisms can have merits-based underinclusive effects, then it seems likely that some of the non-death sentences in those boxes are underincluded for that reason. Such underinclusion would result either from a prosecutor exercising merits-based underinclusion in deciding not to seek the death sentence, or from at least one juror doing so in the sentencing decision. (Indeed, some of the seemingly aberrational non- death sentences in boxes N and P can probably be explained this way as well.) I am not alone in believing that such underinclusion is legitimate and desirable: a broad-ranging consensus supports the desirability of merits-based underinclusion in death sentencing. Many state legislators who support the death penalty have included underinclusion mechanisms in their states' death penalty statutes.[106] Further, state legislators have included these mechanisms voluntarily, as the Supreme Court has never even hinted that states are required to adopt such mechanisms.[107] On the other side of the debate, death penalty opponents, who have lost the fight to abolish the death penalty, would surely, as a fallback position, want to have as many underinclusion mechanisms as possible incorporated in any death penalty statute. Thus, I believe that it is hardly open to debate that merits-based underinclusion in death sentencing is a virtue, not a vice.

F. Supreme Court Underinclusiveness Jurisprudence

The Court's death penalty jurisprudence speaks either directly or inferentially to all of the causes of underinclusiveness that I have identified.

1. The Court Has Never Entered the Fray with Respect to Prosecutorial Underinclusion Due to Mundane Causes.

The Court has not wanted to regulate prosecutorial underinclusiveness resulting from mundane causes,[108] perhaps because it recognizes that it has no effective way-short of declaring the death penalty unconstitutional per se-to regulate this sort of prosecutorial behavior. The Court's position is perhaps most apparent in cases relating to claims of "selective prosecution." To prevail in either a capital or noncapital case, a defendant must show discrimination on some illegitimate basis.[109] In effect, this body of doctrine declares that courts should stay out of the hurly-burly of prosecutorial decision making unless the decisionmaking process is permeated by invidious motives.

2. The Court Has Prohibited States from Curtailing Guilt-Determiners' Power to Exercise Merits-Based Underinclusion

The Court also has sought to prevent states from curtailing the power of guilt-determiners to exercise merits-based underinclusion. The key precedent in this area is Beck v. Alabama,[110] in which the Court found a state law that prohibited lesser-included offense instructions unconstitutional:

[W]hen the evidence unquestionably establishes that the defendant is guilty of a serious, violent offense-but leaves some doubt with respect to an element that would justify conviction of a capital offense-the failure to give the jury the "third option" of convicting on a lesser included offense would seem inevitably to enhance the risk of an unwarranted conviction. Such a risk cannot be tolerated in a case in which the defendant's life is at stake.[111]

Giving the guilt-determiner the "third option," of course, provides the opportunity to underinclude. Beck establishes the constitutional necessity that the guilt-determiner be provided with that opportunity.

3. The Court Has Prohibited States from Undermining the Possibility of Merits-Based Sentencer Underinclusion (and, Indirectly, Has Fostered Merits-Based Prosecutorial Underinclusion)

As with guilt-determiner underinclusion, the Court has actively sought to prohibit states from curtailing merits-based sentencer underinclusion. This practice is illustrated by five important lines of authority, starting with the 1968 decision of Witherspoon v. Illinois.[112] In Witherspoon, the Court analyzed an Illinois statute permitting the challenge for cause of any venireperson who stated Sconscientious scruples against capital punishment, or that he is opposed to the same."[113] The Illinois Supreme Court had construed the statute as permitting the prosecution to challenge for cause any venireperson who "might hesitate" to return a death verdict.[114] The U.S. Supreme Court found the Illinois scheme to be unconstitutional:

[A] jury that must choose between life imprisonment and capital punishment can do little more-and must do nothing less-than express the conscience of the community. . . .

If the State had excluded only those prospective jurors who stated in advance of trial that they would not even consider returning a verdict of death, it could argue that the resulting jury was simply 'neutral' with respect to the penalty. But when it swept from the jury all who expressed conscientious or religious scruples against capital punishment and all who opposed it in principle, the State crossed the line of neutrality. In its quest for a jury capable of imposing the death penalty, the State produced a jury uncommonly willing to condemn a man to die.[115]

Witherspoon clearly shows the Court embracing death sentence underinclusion by prohibiting the state from employing a device-exclusion of death-dubious jurors-that would have tended to minimize the chances of underinclusion.

The Woodson v. North Carolina[116]/Roberts v. Louisiana[117] line of cases[118] prohibiting states from employing a mandatory death penalty for certain categories of murders illustrates the same point. The effect of a mandatory sentence is to negate the possibility of underinclusion as to any defendant who falls into that category. The Court's unwillingness to tolerate such mandatory schemes clearly illustrates its dislike for devices that would eliminate the possibility of underinclusion.

The Lockett v. Ohio[119]/Eddings v. Oklahoma[120] doctrine requiring the sentencer to consider all the mitigating evidence offered by the defendant also undermines Steiker and Steiker's conclusion that the Court is not concerned with underinclusion. The purpose of the requirement is to give the sentencer every possible basis to underinclude a particular defendant. A lesser-known case in this line is instructive. In Mills v. Maryland,[121] the jury instructions and verdict form could have led the jurors to believe that once they found an aggravating circumstance, they had to unanimously find a mitigating circumstance to return a non-death sentence.[122] The Court found this unconstitutional-if any one juror found any mitigating circumstance, the juror had to be permitted to balance it against the aggravating circumstance(s).[123] Mills shows the Court, once again, taking a stand against state practices that would have the effect of minimizing the possibility of underinclusion.

In Caldwell v. Mississippi,[124] the prosecutor and the trial court suggested to the sentencing jury that the ultimate responsibility for imposition of the death sentence rested not with the jury but with the appellate court, because the jury's decision was reviewable.[125] The Court reversed the death sentence:

This Court has always premised its capital punishment decisions on the assumption that a capital sentencing jury recognizes the gravity of its task and proceeds with the appropriate awareness of its "truly awesome responsibility." In this case the State sought to minimize the jury's sense of responsibility for determining the appropriateness of death.[126]
The jury should be forced to "confront and examine the individuality of the defendant,"[127] including particularly those "compassionate or mitigating factors stemming from the diverse frailties of humankind."[128] Arising out of this confrontation would be the possibility of underinclusion-that at least one of the jurors would find something in the defendant worth sparing.

These four lines of authority, taken together, constitute strong evidence that the Court has consistently viewed merits-based sentencer underinclusion as a virtue, and has taken steps to prohibit states from minimizing it. However, it is the fifth line of authority, embodied in McCleskey v. Kemp,[129] that is the real nail in the coffin of Steiker and Steiker's argument that the Court has viewed all underinclusion, whatever its source, as a vice. Steiker and Steiker discuss McCleskey under the fourth of their concerns, heightened procedural regularity.[130] However, McCleskey is much more pertinent to underinclusion because it is the only important case in which the Court has ever been starkly confronted with the opportunity to view underinclusion as a vice calling for a constitutional remedy.

Although not the most famous portion of the opinion,[131] the Court did consider whether McCleskey's death sentence could stand in view of the fact that others with similar culpability had not been sentenced to death.[132] In essence, the Court considered whether underinclusion rendered the death sentence unconstitutional. The Court declined to find a constitutional defect: "[A]bsent a showing that the Georgia capital punishment system operates in an arbitrary and capricious manner, McCleskey cannot prove a constitutional violation by demonstrating that other defendants who may be similarly situated did not receive the death penalty."[133] The Court deemed that it had rejected McCleskey's underinclusion contention way back in Gregg v. Georgia, where it had said:

Nothing in any of our cases suggests that the decision to afford an individual defendant mercy violates the Constitution. Furman held only that, in order to minimize the risk that the death penalty would be imposed on a capriciously selected group of offenders, the decision to impose it had to be guided by standards so that the sentencing authority would focus on the particularized circumstances of the crime and the defendant.[134]
The McCleskey Court concluded its discussion of this point as follows:

Because McCleskey's sentence was imposed under Georgia sentencing procedures that focus discretion "on the particularized nature of the crime and the particularized characteristics of the individual defendant," we lawfully may presume that McCleskey's death sentence was not "wantonly and freakishly" imposed, and thus that the sentence is not disproportionate within any recognized meaning under the Eighth Amendment.[135]
A clearer rejection of the notion that the Court believes capital underinclusion, regardless of its source, to be unconstitutional can hardly be imagined.

These lines of authority not only serve to prohibit states from minimizing the opportunity for merits-based sentencer underinclusion, but also indirectly foster merits-based underinclusion by prosecutors at both the charging and sentencing stages. Tender-hearted prosecutors can defend decisions not to seek death sentences by saying, "While the case may have warranted it, I wasn't completely sure," or, more likely, "I didn't think the sentencer would go for death, so why spend all that extra money?" Even hard-nosed prosecutors can't ignore fiscal concerns, nor do they like to be perceived as having "lost" a case where they sought a death sentence, but the sentencer returned a nondeath verdict. Thus, maintaining the power of the sentencer to underinclude likely has a "trickle-down" effect on prosecutors.

4. The Court's Real and Only Underinclusion Concern: Invidious Underinclusion

We have seen that the Court has not been concerned with prosecutorial underinclusion due to everyday reasons, has sought to preserve the power of guilt-determiners to exercise merits-based underinclusion, and has sought to preserve the power of sentencers to exercise merits-based underinclusion (which likely has a trickle-down effect on prosecutors' exercise of merits-based underinclusion). What is left in the underinclusion grab-bag is, of course, invidious underinclusion at four possible junctures: by prosecutors in charging, by prosecutors in deciding whether to seek a death sentence, by guilt-determiners, and by sentencers. Despite the Court's broad pronouncements that state death penalty systems are required to operate evenhandedly,[136] by process of elimination it is clear that invidious underinclusion is the only sort of underinclusion with which the Court has ever really been concerned.

In understanding the Court's concern with evenhandedness, it is important to recognize the racially charged milieu in which members of the Court made their initial comments about the subject back in Furman. For as many years as they had been on the Court, these justices had seen a steady stream of death penalty cases from the South in which (1) African-Americans were seemingly disproportionately represented as defendants; and (2) many of the defendants had been sentenced to death for rape of white women.[137] In this milieu, it seems quite likely that judicial comments about evenhandedness are in fact coded references to concerns about invidious racial discrimination against African-American defendants.

Fifteen years after Furman, the McCleskey Court dealt invidious discrimination claims in capital cases a near-mortal blow. There, the Court was presented with painstakingly thorough research from a BWP study showing that, all other things being equal, "a defendant's odds of receiving a death sentence are increased by a factor of 4.3 when the victim is white."[138] The data constituted evidence of either overinclusion (killers of whites who were not among the "worst" were being sentenced to death on an invidious basis), underinclusion (killers of blacks who were among the "worst" were not being sentenced to death on an invidious basis when comparable killers of whites were), or, more likely, both. This was about the most compelling scenario that could have been conceived-short of race-of-defendant disparities-to induce the Court to condemn a capital punishment system on the basis of invidious racial discrimination. Nevertheless, the argument only induced four justices to that conclusion. To the majority, McCleskey's proof of invidious discrimination simply was not strong enough.[139] While this was a great defeat for death penalty opponents, the Court did in fact adhere to the position that invidious discrimination, if proven, would be unconstitutional.[140] The most charitable interpretation of the dubious holding that the proof was insufficient is that the Court may have been worried that a contrary conclusion would send the states "back to the drawing board" to come up with systems having some internal review mechanism promoting more consistency. This would, of course, have the concomitant effect of leaving less room for merits-based underinclusion. Perhaps this concern was in Justice Powell's mind when he wrote for the majority:

Discretion in the criminal justice system offers substantial benefits to the criminal defendant. Not only can a jury decline to impose the death sentence, it can decline to convict or choose to convict of a lesser offense. Whereas decisions against a defendant's interest may be reversed by the trial judge or on appeal, these discretionary exercises of leniency are final and unreviewable. Similarly, the capacity of prosecutorial discretion to provide individualized justice is "firmly entrenched in American law." As we have noted, a prosecutor can decline to charge, offer a plea bargain, or decline to seek a death sentence in any particular case. Of course, "the power to be lenient [also] is the power to discriminate," but a capital punishment system that did not allow for discretionary acts of leniency "would be totally alien to our notions of criminal justice."[141]
G. The Court's Primary Concern: Minimizing Overinclusion

I have demonstrated that the Court's dual concerns with respect to underinclusion have been to maximize the opportunities for the merits-based variety, while minimizing the instances of the invidious kind.[142] This is consistent with the Court's primary goal of minimizing overinclusion. The Court has taken a broad view of overinclusion, insisting that "worstness" must be determined by considering not only the facts about the crime, but also any ameliorating facts about the defendant.[143] Thus, the Court's view of overinclusion encompasses Steiker and Steiker's individualization concern.[144] It also encompasses the concern for heightened procedural reliability[145] because there must be some process ensuring the sentencer's opportunity to hear all the relevant information.

This overinclusion concern has dictated the requirements that the Court has imposed on the states (which are, as Steiker and Steiker correctly point out, relatively minimal[146]): some specification of what makes a murder among the "worst;"[147] a virtually unlimited opportunity for the defendant to introduce mitigating evidence;[148] and a sentencing process that focuses the sentencer's attention on the seriousness of the decision.[149] While the seeds of this regulatory regime were planted in Furman, and sprouted in Gregg and its companion cases, I think the true shape of the mature plant only became clear in Lockett, where the Court made it crystal clear that states could not erect any barrier to the presentation of mitigating evidence.[150]

To summarize my argument so far, the academic underinclusionists are wrong in identifying one of the Court's goals as rectifying states' "failure to treat equally deserving cases alike" (again, except for invidious underinclusion). Rather, the Court has tried to preserve the power of prosecutors and sentencers to underinclude in potential capital cases. In the following Part, I will argue that not only have Steiker and Steiker misidentified one of the Court's goals, but they also have not proven their claim that the Court's jurisprudence has failed to achieve any measurable degree of success as to its chosen goals. Indeed, the available evidence points in the opposite direction.

III. STEIKER AND STEIKER HAVE NOT PROVEN THAT THE COURT'S REGULATORY EFFORT HAS BEEN AN INEFFECTIVE DISASTER AND HAVE IGNORED EVIDENCE TO THE CONTRARY

Steiker and Steiker's argument for the ineffectiveness of the Court's capital punishment jurisprudence has both a premise and a conclusion. The premise is that the Court has not imposed significant requirements or prohibitions[151] mandating states to create systems having stricter standards than pre-Furman systems,[152] but, through the Lockett/Eddings doctrine, has actually forced states back toward pre-Furman-like systems.[153] This means that while present systems appear to be more nonarbitrarily selective and provide for more structured and informed decisionmaking, they have virtually the same potential to operate in a systematically arbitrary fashion.[154] I believe this premise is flawed because there are at least two significant ways that current systems not only appear to provide for less arbitrary decision making, but actually do so: the required statutory narrowing of death-eligibility, and sentencing hearings that permit the defendant to present mitigating evidence. I will elaborate on this point in subpart A, below.[155]

The conclusion of Steiker and Steiker's syllogism is that because the Court's jurisprudence leaves virtually the same potential for arbitrariness as in pre- Furman systems,[156] the jurisprudence has been "a disaster, an enormous regulatory effort with almost no rationalizing effect."[157] Even assuming, arguendo, that their premise is correct, Steiker and Steiker's conclusion is flawed. The Steikers' argument consistently leaps from the could be to the is: because current state systems, minimally regulated by the Court, invite arbitrary application,[158] the Court's jurisprudence is an ineffective disaster.[159] Steiker and Steiker miss an indispensable logical step: they have to prove that systems with virtually the same potential to operate in systematically arbitrary fashions are actually living down to that potential. The authors do not even attempt such proof, and in fact, they ignore important evidence to the contrary. I will explain this point in subpart B, below.[160]

A. The Flaw in Steiker and Steiker's Premise

1. The Winnowing Effect of Statutory Narrowing of Death-Eligibility

Steiker and Steiker minimize the winnowing effect of aggravating circumstances in Georgia,[161] citing BWP's work showing that most post-Furman Georgia murderers are death-eligible because their cases involve at least one aggravating circumstance.[162] While it is true that the aggravating circumstance requirement's winnowing effect is not huge, neither is it de minimis: BWP found that of the 606 jury-trial murder convictions in their post- Furman study, 123 defendants were immune from the death penalty because no aggravating circumstance existed.[163] This is an exclusion rate of twenty percent that did not exist pre- Furman.[164] Admittedly, these excluded defendants would have been unlikely candidates for death even under the pre- Furman system. But that also means that they would have been candidates from which one could expect instances of overinclusion. If, as I have argued, the Court has always been primarily concerned with overinclusion, then the Court should consider it a victory to have forced Georgia to define these defendants out of the death-eligible pool. In this respect, the post-Furman Georgia system is demonstrably more nonarbitrarily selective than the pre-Furman system. While no work equal in scope with BWP's dated Georgia study has been done in any other death penalty jurisdiction, there is no reason to expect that the statutory narrowing of death eligibility in other death penalty states after Furman did not have the same beneficial effect on selectivity.[165]

2. The Effect of Nonperfunctory Sentencing Hearings

Prior to 1970, a death penalty trial in Georgia was unitary, meaning that the jury deliberated on both guilt and sentence at the close of the state's case.[166] Under this procedure, there was no opportunity for the sentencer to have the benefit of mitigating information that might have been developed at a sentencing hearing. In 1970, two years before Furman, the Georgia Legislature mandated bifurcated proceedings in death penalty cases.[167] This required a separate proceeding to determine sentence after the jury reached a decision on guilt.[168] This was a step forward, but BWP, who examined a sample of the Georgia death penalty case files for the time between the institution of the bifurcated proceeding and the time of Furman had this to say about those sentencing proceedings: "[P]rior to Furman, the sentencing phase of Georgia's bifurcated procedure was a perfunctory affair. Normally, the only additional evidence offered during the sentencing hearing was the defendant's prior criminal record, if any."[169] BWP also became intimately familiar with hundreds of prison case files from the five years following Georgia's death penalty statute that was eventually upheld in Gregg. Of sentencing hearings during that time period, BWP had this to say: "After Furman and, perhaps more importantly, after Lockett v. Ohio, which sensitized attorneys to the significance of mitigating circumstances, the evidence offered during the sentencing phase of capital cases in Georgia became much more extensive."[170] Here, then, is another way in which the difference between the pre- and post- Furman worlds is not just a chimera-the sentencer in Georgia got significantly more relevant information in the post-Furman era.[171] Again, there is no cause to believe that Lockett has not had the same effect in other death penalty states, and because the Court views more information for the sentencer as a crucial device for minimizing overinclusion, the Court should be quite pleased with the effects of Lockett.

3. Has the Court Succeeded in Minimizing Overinclusion Due to Race?

The two primary ways in which invidious racial discrimination can infiltrate the system are via race-of-defendant effects-African-Americans being more likely to be sentenced to death for comparable crimes-and race-of-victim effects-killers of African-Americans being less likely to be sentenced to death than those with non-African-American victims. BWP's 1973-1978 study found that race-of-defendant effects had decreased to the point of being a nonfactor.[172] On its face, this finding cuts in favor of the effectiveness of the Court's death penalty jurisprudential efforts. However, BWP also concluded that this decline was not due to the reforms prompted by Furman.[173] Nonetheless, the Court could plausibly declare victory on this point by invoking the aphorism, "A win is a win." However, the statistics in McCleskey suggest that the Court cannot declare victory regarding race-of-victim effects.[174] Thus, the efficacy of the Court's jurisprudence regarding its overinclusion subgoal of minimizing invidious underinclusion is highly debatable.

B. The Flaw in Steiker and Steiker's Conclusion

Despite the demonstrable rationalizing effects of the statutory narrowing of death eligibility and more complete sentencing hearings, Steiker and Steiker correctly argue that the potential for arbitrary overinclusion still exists. After all, no human-judgment-driven system is perfect. Nevertheless, to prove that this potential is being realized to the same extent as in the pre-Furman era, Steiker and Steiker would have to present something more than assumption and conjecture. This they do not do, cryptically citing only one source in one footnote[175] for the general proposition that "a substantial number of death sentences continue to be imposed in a fashion that can only be described as 'freakish.' "[176] Perhaps even worse than their lack of proof is their failure to address significant research that tends to prove exactly the opposite conclusion, not only as to overinclusion (a concern the Court embraces) but also as to underinclusion (a concern the Court embraces only in part).

1. Empirical Evidence of the Decline in Overinclusion

a. BWP's Georgia Study

When it comes to empirical analysis of the real-world effects of a death penalty system, the work of BWP in Georgia[177] towers above all others, and is the only substantial research effort that has compared a pre- Furman system with a post- Furman system.[178] BWP meticulously analyzed 131 pre- and post- Furman Georgia homicide cases, compiled the data, and subjected it to sophisticated statistical analysis. The BWP data, although used by BWP to measure underinclusion,[179] can easily be adapted to examine overinclusion.

BWP sorted 293 death-eligible (i.e., guilty of murder) pre-Furman cases into six culpability levels (with level one being least culpable, and level six being most culpable) based upon numerous factors.[180] They then calculated the death sentencing rate within each level. Their results are set forth in Figure 3.[181] The data is easy to understand: at level one, for example, only one out of 110 defendants was sentenced to death, a rate of .01; while at level six, ten out of ten defendants were sentenced to death, a rate of 1.00.

FIGURE 3

FIGURE 3


A. Case Culpability Level

from 1 (low) to 6 (high)

B. Death-Sentence Rate

1

.01 (1/110)

2

.08 (8/95)

3

.23 (10/43)

4

.53 (8/15)

5

.35 (7/20)

6

1.00 (10/10)




The chart clearly illustrates some overinclusion, although exactly how much depends upon what cutoff one establishes for the dividing line between overinclusion and underinclusion. To me, the .23 rate at level three does not illustrate overinclusion-almost a quarter of the murderers in that level were determined unanimously by juries to be among the "worst" murderers. On the other hand, any defendant in a level where the death sentencing rate falls below .10 (ten percent) has figuratively been "struck by lightning."[182] The data in this chart does not require any thinner parsing: using the .10 cutoff, the proportion of overincluded defendants to the whole 293-person population is .03 (9/293). This rate is quite low in and of itself.[183] However, the proportion of overinclusive death sentences within the category of the death-sentenced is significant-nine of the forty-four death sentences were overinclusive, a proportion of .20 (twenty percent).

BWP performed the same analysis on 606 Georgia murder convictions in the five years following the 1973 enactment of the new Georgia death penalty statute prompted by Furman. The lack of any aggravating circumstance removed 123 defendants from death penalty eligibility,[184] which left 483 eligible cases. The BWP figures for those cases are shown in Figure 4.[185]FIGURE 4

FIGURE 4


A. Case Culpability Level

from 1 (low) to 6 (high)

B. Death-Sentence Rate

1

.02 (6/276)

2

.14 (9/65)

3

.38 (18/47)

4

.65 (22/34)

5

.85 (23/27)

6

1.00 (34/34)




Here, more discrimination concerning where to draw the cutoff for overinclusion is appropriate. Specifically, were the death-sentenced defendants at level two, where the rate is .14, overincluded? If we adopt a .10 cutoff, thereby deeming level two's death-sentenced defendants to not have been overincluded, then there is a significant improvement from the pre-Furman data: only six of the 606 defendants (remember, the 123 defendants who were convicted of first-degree murder but were not death-eligible because of the lack of an aggravating circumstance have to be included to make a valid comparison, because they would have been death-eligible pre-Furman) were overincluded, a proportion of .001 (compared to the pre-Furman proportion of .03); and only six of the 112 death sentences were overinclusive, a proportion of .05 (compared to the pre-Furman proportion of .20). If we draw the cutoff higher, say at .15 (which would be reasonable), or .20 (which would be too high, but nevertheless would make no difference with these data, because the .14 rate at level two jumps to .38 at level three), the improvement still exists, although it becomes less significant. There also are fifteen defendants out of the 606-person population who were overincluded, a proportion of .025. While that improvement may be a modest one compared with the pre-Furman proportion of .03, such modesty could be because there really was not a great deal of improvement that could be made-the three-percent pre-Furman overinclusion rate may be about as close to "perfect" operation as any human-judgment-driven system is likely to get.[186] However, another comparison shows more significant improvement in the post- Furman era-the overincluded defendants as a proportion of the death-sentenced defendants is .13 (fifteen of 112), rather than the pre-Furman .20 rate.

Whether these pre-Furman to post- Furman improvements exist in states other than Georgia has not been systematically studied. Nor has Georgia data since 1978 been examined.[187] However, the BWP study does provide important evidence to counter Steiker and Steiker's conclusion in one of the most active death penalty states during a critical time period. The BWP evidence also seems to provide the basis for surmising that the same improvements may well have occurred in other death penalty states, and for other time periods up to the present.

b. Some New Evidence That Overinclusion Is Not a Widespread Defect in Georgia's System

To determine whether BWP's findings about the decrease in overinclusion in Georgia were an artifact of the 1970s, I analyzed more recent cases in Georgia in which defendants received death sentences. I chose the twenty-five cases most recently decided by the Georgia Supreme Court on direct appeal.[188] This sample includes almost one-quarter of the inmates on Georgia's death row,[189] and covers Georgia Supreme Court opinions from early 1988 through late 1995. Although this research is based upon court opinions alone and does not approach the thoroughness of the BWP study, it is nonetheless illuminating.

I begin with a general comment: the facts in every one of the twenty-five cases (except one[190]) serve immediate notice that the case is something more than a run-of-the-mill homicide. Six recurring exacerbating motifs (hereinafter "exacerbators") are apparent. By far the most commonly occurring (fifteen cases) is what I will call "overkill": the method of killing is particularly repulsive because the defendant inflicted more damage than was necessary to cause death. Four cases involve more than fifteen stab wounds;[191] three involve multiple gunshot wounds inflicted in other than a continuous burst;[192] two involve repeated bludgeoning;[193] one involves both ligature and manual strangulation;[194] and four involve a potpourri of methods: (1) shooting, bludgeoning with a paint can, and beating with a crowbar;[195] (2) beating with a baseball bat, shooting, and dismembering (although the victim may have been dead at the time of dismembering);[196] (3) stabbing over a hundred times and then bludgeoning in the head with a hammer;[197] and (4) stabbing and then beating in the head with a clothing iron.[198]

After overkill, the second exacerbator is kidnap/rape/murder (four cases).[199] The third is multiple homicide (six cases).[200] The fourth is that the defendant's prior criminal record clearly indicated that he had not learned his lesson (three cases): (1) the defendant was out on parole from a sentence for a prior murder;[201] (2) the defendant committed the homicide while in prison for a capital crime;[202] and (3) the defendant committed the kidnap/rape/murder on the same day he was released on parole.[203] The fifth exacerbator is that defendant was on a crime spree in which he committed serious crimes other than the homicide that resulted in the death sentence (seven cases): (1) the defendant was involved in a series of motel-patron robberies, and had shot a person in another robbery;[204] (2) the defendant had shot a person two hours earlier and was pistol-whipping another person when interrupted by the murder victim;[205] (3) the defendant planned and executed a series of rapes;[206] (4) after killing the victim, the defendant drove to another state and killed again;[207] (5) after bludgeoning the victim with a hammer and stealing his car and some personal belongings, defendant drove to another state where he robbed, bludgeoned, and shot another person;[208] (6) defendant shot two other people during a robbery in addition to the two who died;[209] and (7) defendant tied up, robbed, and seriously battered another victim in addition to the one who died.[210] The sixth exacerbator is that the defendant killed a child (two cases).[211] There is only one[212] of the twenty-five cases that does not contain at least one of these exacerbating motifs, and many cases involve more than one, as is shown in the chart below.[213]

The exacerbators are, of course, only a rough-and-ready indicator that these death sentences were not overinclusive. A more illuminating test would be to try to estimate what percentage of similar cases resulted in death sentences. I analyzed the cases in two ways. First, I applied the Barnett method, which has already been explained.[214] Second, I used a method developed by BWP[215] that categorizes cases based upon the following factors: the number of persons killed by defendant;[216] whether the defendant committed a serious contemporaneous offense (kidnapping, rape, or armed robbery);[217] whether there were serious aggravating circumstances;[218] whether defendant had a prior record;[219] whether there were mitigating circumstances;[220] and whether there were minor aggravating circumstances.[221] The BWP method is contained in Appendices C and D. The two methods are complementary in that each considers significant factors that the other does not, or uses the factors differently in the analysis. Thus, it would not be unusual for cases to end up in boxes in the respective charts that had significantly different death-sentencing ratios. I considered a case worthy of further discussion in terms of overinclusion if its ratio did not rise to the level of .30 using either method; i.e., less than thirty percent of defendants in the boxes in which the defendant's case fell in the respective charts were sentenced to death. In using these two methods, I made the large but commonsensical assumption that the factors about cases which were significant in the 1970s remained constant in the late 1980s and early 1990s.[222] The results of my analyses appear in summary form in Figure 5.

There are four cases-all involving intrafamily killings-that require closer examination for overinclusion because they do not rise to the .30 ratio under either the Barnett or BWP criteria: cases 7, 13, 15, and 18. I will examine these cases in order from the easiest to the hardest to understand why a death sentence was imposed.

Case 15 is the lowest scoring of all the cases-garnering only a .01 ratio in the Barnett analysis, and a .00 under the BWP analysis-yet it is the easiest one to understand why a death sentence was imposed. What the case really shows are the limitations of both the Barnett and BWP analyses. In this case, the defendant's ten-year-old son annoyed him by playing with a toy tractor after the defendant asked him to stop.[223] Defendant found his shotgun (which the family had hidden), and shot the boy while the child begged for his life.[224] Defendant then stated, "I couldn't learn him nothing by beating him with a belt, so I guess I learned him something this time."[225] The defendant was, though, quite drunk at the time.[226] The case scores only 1, 0, 1 under the Barnett criteria because, as to the first Barnett factor, the case does not meet any of the criteria for raising the presumptive 1 to a 2; as to the second factor, the defendant knew the victim, resulting in a score of 0; and as to the third factor, while there was no claim of self-defense, thus giving a score of 1, the case does not meet any of the criteria for the vileness sub-inquiry concerning the horrific nature of the killing that would raise the 1 to a 2. The case scores low under the BWP criteria because the only factor

FIGURE 5

FIGURE 5


CASE #

BARNETT RATIO

BWP RATIO

EXACERBATORS

1

.28

.68

Overkill; kidnap/rape; child victim

2

.56

.32

Double homicide

3

.81

.60

Overkill

4

--[227]

.25

Prior record; spree

5

.56

.03

Overkill

6

.28

.60

Overkill

7

.01

.00?[228]

--

8

.28

.32

Double homicide

9

.56

--[229]

Overkill

10

.81

.69

Overkill; spree

11

.86

.69

Kidnap/rape; spree

12

.28

.89

Overkill

13

.28

.03

Overkill

14

.56

.69

Overkill: spree

15

.01

.00

Child victim

16

.28

.60

Overkill; spree

17

.81

.60

Overkill

18

.05

.03

Overkill

19

.28

.88

Double homicide

20

.30

.88

Double homicide; spree

21

.81

.25

Kidnap/rape; prior record

22

.28

.32

Triple homicide; 2 child (?) victims

23

.56

.42

Overkill

24

.28

.32

Double homicide

25

.81

.69

Overkill; spree



working against the defendant is a serious aggravating circumstance (victim pled for life). The case did not involve a serious contemporaneous felony, the defendant had no prior record, the defendant presented the mitigator of drunkenness, and there was no minor aggravating circumstance. Neither analysis captures the significance of the victim being a child, the trivialness of the "provocation," or the defendant's subsequent boasting about the homicide. Common sense indicates that this is actually quite an aggravated case, and that many-if not most-juries would impose death sentences in similar cases.

The death sentence in case 13 is the next easiest to understand. In this case, the defendant stabbed his sleeping wife over a hundred times, then bludgeoned her in the head with a hammer.[230] His defense was that he was laboring under the delusion that his wife and another man were plotting to kill him,[231] but the opinion presents no evidence that defendant had a history of mental problems. Clearly, overkill explains this death sentence. The Barnett ratio of .28 does not raise a real problem here-while that ratio is certainly only moderate, box M, into which this case falls, is a popular one. It contains the third highest number of cases of any of the Barnett's boxes, and accounts for eleven death sentences in the cases Barnett studied. As to the BWP criteria, the case scored a dismal .03 because it did not involve a serious contemporaneous offense, the defendant had no prior record, and the defendant presented a mitigating circumstance of believing he had a moral justification. The explanation for this death sentence may be that the sentencers in the actual cases may not have taken into account the defendants' mitigating circumstances. It is notoriously difficult to convince sentencers that evidence of mental problems is strong enough to operate as a mitigating factor.[232] It is very likely that the sentencer here rejected the mitigating evidence, which would mean that the case actually falls into BWP box B12, where the ratio is .42. Or perhaps this case simply shows the limits of both the Barnett and BWP modes of analysis to capture the nuances of every case-the overkill here seems commonsensically to lead to the conclusion that the case is not overincluded.

Case 18 is almost a carbon copy of case 13, with one important exception. In case 18, on the day defendant's wife obtained a good-behavior warrant, defendant (with his wife's agreement) went to their apartment to gather some belongings.[233] While there he stabbed his wife numerous times, including a neck slash that almost decapitated her.[234] The one important difference from case 13 is that here the defendant had a documented history of psychological and substance abuse problems.[235] As in case 13, the death sentence is certainly a result of the defendant's overkill. Case 18 scores lower than case 13 using the Barnett analysis because the defendant's documented mental health problems lower the score on the first factor from 1 to 0. The case scores the same as case 13 under BWP criteria for the same reasons: no serious contemporary offense, no prior record, and a mitigating circumstance. The most likely explanation that would save this sentence from being classified as overinclusive is the same as in case 13: the sentencer was not convinced by the mental health testimony. Indeed, the opinion points out that the prosecutor presented opposing experts to testify that the defendant was a malingerer.[236] If the jury believed that testimony, the case would fall into another box with a ratio of .42.

The hardest case in which to explain the death sentence as not overinclusive is number 7. In this case, the defendant broke into his ex-wife's house armed with a gun.[237] When she and her son arrived home, the son tried to shoot the defendant, but the gun wouldn't fire.[238] Defendant hit his ex-wife at least fifty times, then dragged her outside, pulled her head back by her hair, and shot her in the forehead.[239] This case is the only one of the twenty-five that does not involve an exacerbator. Further, the case's Barnett score of 1, 0, 1 places it in a category where only one of 106 cases resulted in a death sentence. Under BWP analysis, it falls into a box where the ratio is .00 (unless this qualifies as an "execution-style" killing, in which case it moves to a box with a .42 ratio). What all this shows is that despite the appalling nature of the case, it is a fairly typical domestic killing. The other twenty-four cases demonstrate that death sentences are not imposed in domestic killings unless there are multiple homicides (as in cases 2, 22, and 24), overkill (as in cases 12, 13, 14, 18 and 23), or a child victim (as in case 15). None of these factors are present here. Although the case does involve an armed burglary, and an arguably cold, execution-style slaying, it may represent an instance of overinclusion.

My conclusion from analyzing these cases is that, with one possible exception,[240] the Georgia system has not been overinclusive over the past six years. This conclusion accords with the very low rate of overinclusion illustrated by the BWP study of 1973-1978.[241] Even though neither BWP nor I examined cases from 1979-1988 for overinclusion, I feel fairly confident in asserting that Georgia's post-Furman system has not exhibited a significant degree of overinclusion. Whether the same can be said for other states is an open question, but perhaps the Georgia data provides hope that what is true in Georgia is also true elsewhere.

2. Empirical Evidence of the Decline in Underinclusion

BWP specifically set out to compare underinclusion in the pre- and post- Furman cases.[242] Their standard for judging underinclusion was the following: "[W]e classify death sentences as presumptively excessive [the BWP synonym for underinclusive] if the frequency with which death sentences occur in other cases classified as 'similar' by our various measures is less than .35. If the death-sentencing rate among similar cases exceeds .80, we classify a death-sentence case as presumptively evenhanded."[243] Using an "overall culpability" index based upon seventeen legitimate case characteristics[244] and regression analysis,[245] BWP studied a sample of pre- and post-Furman cases. This comparison of the two sets of cases showed the post-Furman system to be significantly less underinclusive and more evenhanded than the pre-Furman system:

[I]n comparison to [the pre- Furman figure], there is more selectivity in the post-Furman system. For example, only 29 percent of the post- Furman death-sentence cases possessed culpability scores equal to or less than the culpability score of the 95th percentile life-sentence cases, a decline from the comparable pre-Furman figure of 61 percent . . . . Only 13 percent of the post-Furman death sentences appear in categories of cases with death-sentencing rates of less than .35, and more than half . . . of the post- Furman death sentences occurred in cases for which the death-sentencing rate among similar cases exceeds .80.

The resulting proportion of presumptively excessive sentences has dropped from .43 (19/44) pre-Furman to .13 (15/112) post-Furman, and the proportion of [presumptively] evenhanded sentences has grown from .23 (10/44) pre-Furman to .51 (57/112) post-Furman. This is evidence of increased selectivity.[246]

One also should note from the two BWP charts above[247] that the death-sentencing rate rose in the three crucial midlevel categories that give rise to the underinclusion concern-from .23 to .38 in level three; from .53 to .65 in level four; and from .35 to .85 in level five-indeed, the level five proportion moved from the brink of presumptive excessiveness to above the threshold for presumptive evenhandedness. Steiker and Steiker fail to acknowledge the existence of this evidence; indeed, they cite the very BWP book in which this data appears for the proposition that "critics" argue that the Court has "turned its back on regulating the death penalty and no longer even attempts to meet the concerns about arbitrary . . . imposition of death that animated its 'constitutionalization' of capital punishment in Furman."[248]

Whether we should be heartened by this decline in underinclusion is another question. A decline in merits-based underinclusion would be a mixed blessing: while it would make the system more evenhanded, it also would tend to show that the system is not erring as much as it perhaps should on the side of not imposing death in debatable cases. A decline in mundane underinclusion would be neutral. A decline in invidious underinclusion would be welcome, but there is no significant evidence of such a decline due to anything the Court has done.

C. Are the Post- Furman Declines in Overinclusion and Underinclusion Attributable to the Court's Jurisprudence?

A last hope for Steiker and Steiker's conclusion might be that, even assuming these improvements in overinclusion and underinclusion exist, they were not a result of the Court's death penalty regulatory jurisprudence, but instead resulted from some other factor(s). Unfortunately for Steiker and Steiker, BWP considered this possibility, and rejected the conclusion that the Court's efforts had not played a significant role in the process.[249]

BWP noted that improvements in the system began immediately and abruptly after the Georgia Legislature retooled the death penalty statutes in the wake of Furman:

The information . . . suggests that Georgia's 1973 statutory reforms caused the observed differences in Georgia's pre- and post-Furman capital-sentencing patterns. The data for 1974, the first post- Furman year with a large number of death sentences, indicate a considerably greater degree of selectivity compared to the pre-Furman period. The trend lines . . . which indicate the yearly median-culpability scores for life- and death-sentence cases, also suggest that the impact of the post- Furman procedural changes was, except in 1976, a sustained one. The data . . . which presents the average rather than the median culpability scores for the life- and death-sentence cases from 1970 to 1978, suggest an even greater increase in selectivity post- Furman.[250]
BWP hypothesized four changes other than the statutory reforms prompted by Furman that might have accounted for the improvements. These include better record-keeping in the state system;[251] a general trend toward consistency and reasonableness on the parts of prosecutors and juries, perhaps from better education and improved living standards;[252] initially greater selectivity by prosecutors and judges due to the fear that the Court would exercise close supervision over the new system;[253] and a general decline in racial discrimination.[254] While believing that each of these alternative hypotheses probably had some explanatory power,[255] BWP concluded that the impact of the statutory reforms was nonetheless important: "We remain persuaded, therefore, that Georgia's 1973 statutory reforms contributed significantly to the decline of excessiveness that we observe in our post-Furman data."[256] Those statutory reforms were, of course, a direct product of the Court's death penalty jurisprudence. This hardly supports Steiker and Steiker's conclusion that the Court's efforts-judged through the lens of its own concerns-have been an ineffective disaster in Georgia. The Georgia data gives reason to suspect that the same ameliorative effect has occurred in other states.

IV. CONCLUSION

Steiker and Steiker begin the second paragraph of their article with the observation, "Virtually no one thinks that the constitutional regulation of capital punishment has been a success."[257] If the authors had said "complete success," surely no one, including the Supreme Court justices, would disagree. But the Court would not-and should not-agree that its efforts have been an ineffective disaster. The best available evidence shows that the Court's regulatory death penalty jurisprudence has been successful in decreasing overinclusion, which is the primary vice that the Court has seen in death penalty systems for the last quarter of a century. One can argue that the Court was wrong to focus (or focus so exclusively) on minimizing overinclusion, or that the Court could have devised better ways to achieve that goal.[258] But let's give credit where credit is due-the populations of death rows since 1972 very likely comprise a more carefully selected and "worse" collection of malefactors than before 1972. This is not an insignificant achievement. So, while death penalty opponents rue the Court's failure to regulate death more severely, I doubt that there is one experienced capital defense lawyer in this country who would rather return to the pre-Furman era. Perhaps not even many prosecutors would want to return to the days before Furman, when such an unguided power of life and death rested in their hands.

APPENDIX A[259]

CLASSIFICATION RULES

I. THE CERTAINTY THE DEFENDANT IS A DELIBERATE KILLER

Score the case either 0, 1, 2 on this dimension, applying the following criteria:

(i) The case is rated 0 if any of the following circumstances pertain:

(1) The narrative indicates the evidence in the case seemed weak (e.g., "case based solely on circumstantial evidence").

(2) The narrative mentions evidence that worked against the view that the defendant was guilty (e.g., tests for residue on the defendant's hand from firing a gun were negative).

(3) It seems clear that the defendant neither ordered the killing nor was the trigger man. (Note that (3) differs from the weaker statement that it is uncertain whether the defendant was the trigger man.)

(4) The killing has an "accidental" touch about it, because

(a) a fairly long period (perhaps a week or more) elapsed between the incident and the victim's death, or

(b) the death was caused by a shot fired somewhat randomly (e.g., through a door), or

(c) the death was caused by a beating similar to previous beatings of the victim by the defendant.

(5) There is reason to doubt that the defendant's actions in themselves would have caused the victim's death (e.g., (i) the defendant beat the victim, but it was a co-perpetrator's stabbing that killed him, or (ii) the defendant's beating of the victim induced a heart seizure).

(6) The defendant was one of several participants in a conspiracy to kill, but took no part in the actual killing.

(7) The narrative mentions that the defendant was previously treated for mental problems (e.g., institutionalized). Neglect references to insanity if the defendant has no apparent medical history.

(ii) The case is rated 2 if any of the following elements were present:

(1) The killing was a murder-for-hire, and the defendant was either the sole instigator or the executioner.

(2) The defendant plotted to kill the victim (e.g., a wife and her lover arrange to murder her husband). If, however, the defendant was one of several plotters, and clearly not the actual killer, assume (2) is not satisfied.

(3) The narrative mentions that the defendant was officially implicated in other killings.

(4) The narrative mentions that the defendant had tried previously to kill the victim.

(5) The defendant announced in advance to a third party an intention to kill the victim. (Neglect this condition in a lover's triangle or lover's quarrel case, or when the third party was a co-perpetrator.)

(iii) If the killing warrants neither a 0 nor a 2, give the case a rating of 1. If the killing satisfies conditions for both 0 and 2, also rate it 1. Most "common" slayings, such as killings during armed robberies or during barroom fights, would warrant this intermediate classification. Indeed, a 2 reflects unusually clear evidence of premeditation, while a 0 reflects unusually large doubt that the defendant knowingly acted to cause the victim's demise.

II. THE STATUS OF THE VICTIM

On this dimension, the score is either 0 or 1.

Give a score of 0 if:

(1) The victim was a relative of the defendant (even his or her child).

(2) The victim was a friend of the defendant. (Interpret the word "friend" loosely; if, for example, two people of similar age are riding together voluntarily in a car, consider them friends. However, the mere fact that two people know each other is not sufficient. Neighbors of vastly different ages, or the bank teller and the depositor, are not assumed friends barring other evidence of social ties.)

(3) The victim was an enemy of the defendant, though not the defendant's employer. (Interpret the word "enemy" loosely; if, for instance, the victim and defendant vied for the affections of the same woman, if the victim had harassed one of the defendant's loved ones, if there was a feud of some sort that turned violent, assume enmity existed. If, however, the victim could be viewed as the defendant's employer - whether as (say) his supervisor in a factory or the person who hired him to perform some chores - do not give a score of 0 under (3).)

(4) The victim, although a stranger to the defendant, acted in a highly provocative manner just prior to the killing (e.g., racial taunts).

(5) The victim was engaged in an illegal or often-disapproved activity at the time of the killing (e.g., a drug dealer, a prostitute or prostitute's customer, owner of a homosexual bathhouse, etc.).

If the case does not warrant the rating 0, give it the score 1. 1 is the appropriate rating for most stranger-to-stranger killings and those in which the defendant only knew the victim in the latter's official capacity (e.g., as employer, or attendant in a local gas station). If there are several victims, give the case a 0 if any of those slain qualify for it.

III. THE HEINOUSNESS OF THE MURDER

There are two aspects to this dimension: the question whether self-defense motivated the killing and how "gruesome" it was.

Self-defense is an element in the case under any of the following circumstances:

(1) The victim had at hand a deadly weapon at the time of the killing. Merely having a gun in the store or house does not satisfy (1).

(2) The victim was killed with his own weapon. (This is taken to imply (1) is satisfied even if the narrative does not explicitly say so.)

NOTE: If the victim was a police officer, do not invoke self-defense in (1) or (2) unless the officer fired shots before the defendant did.

(3) The victim had threatened to kill the defendant or one of the defendant's loved ones.

(4) The victim had attacked the defendant at the time of the killing.

If none of the above conditions existed, self-defense was not a mitigating circumstance in the homicide.

NOTE: If the only evidence for self-defense is the defendant's uncorroborated claim, assume its absence even if any of (1)-(4) is alleged.

A homicide is classified as vile if one of the following circumstances is present:

(1) It was accompanied by rape, or sexual abuse, either against the victim or someone in the company of the victim.

(2) There were at least two homicide victims.

(3) The deceased was a kidnapping victim at the time he was slain.

(4) Psychological torture preceded the killing (e.g., Russian roulette, a sustained period of terror).

(5) The victim was shot several times in the head at close range.

(6) The killing was execution-style (i.e., victim forced to kneel or squat, then shot in head).

(7) The death was caused by strangulation, or arson.

(8) The death was caused by a drowning in which physical force kept the victim below water.

(9) The killing involved ten (10) or more shots or stab wounds, except when the murder weapon was a penknife or other small cutting instrument.

(10) The physical details of the killing are unusually repulsive (e.g., the victim drowned in his own blood).

(11) The body was mutilated, or otherwise grossly disfigured (except in an attempt to conceal the homicide).

(12) The killing was performed with a bizarre weapon (e.g., a hacksaw, a claw hammer, an ice pick).

(13) The defendant apparently derived pleasure from the very act of killing. (This is distinct from his believing the victim deserved to die, and taking pleasure on that account.)

(14) The crime was specifically described in the narrative as extremely bloody.

Absent all these circumstances, the homicide is categorized as not vile. Despite the length of the list above, most "simple" shootings, stabbing, and beatings would not be classified as vile under these rules.

APPENDIX B

APPENDIX B


CASE CITATION &

DATE OF DECISION

SALIENT FACTORS

BARNETT CRITERIA

BWP CRITERIA

1. Wellons v. State,

463 S.E.2d 868

(Ga. 1995)

11-20-95

D kidnapped and raped 15-year-old girl, then strangled her by hand and with cord. D had been drinking. D claimed insanity.

1, 0, 2

.28

B26

.68

2. McMichen v. State,

458 S.E.2d 833

(Ga. 1995)

7-14-95

D shot and killed estranged wife and her boyfriend after prolonged drinking and arguing, then walked his 5-year-old daughter past the bodies and left her in a truck within view of scene. D showed immediate remorse. D had history of depression

2, 0, 2

.56

C1

.32

3. Crowe v. State,

458 S.E.2d 799

(Ga. 1995)

6-26-95

During armed robbery, D shot victim in back, then bludgeoned with paint can and crowbar before victim died.

1, 1, 2

.81

B27

.60

4. Burgess v. State,

450 S.E.2d 680

(Ga. 1994)

12-5-94

During armed robbery, D shot victim who failed to obey order to remove hands from pockets. Several days earlier D had shot another person in another armed robbery. D involved in at least two other armed robberies. D on parole from life sentence for earlier murder. D denied being trigger man.

2, 1, 1

(No Barnett cases to form ratio, but ratios in other categories where three digits total to 4 are .81 and .56)

B24

.25

5. Hittson v. State,

449 S.E.2d 586

(Ga. 1994)

10-31-94

D, at behest of cohort, hit dozing victim over head several times with ball bat, then shot victim in head. Body was dismembered --whether victim was still alive at time unascertainable by coroner. D had been drinking. D claimed insanity.

2, 0, 2

.56

B9

.03

6. Ledford v. State,

439 S.E.2d 917

(Ga. 1994)

2-21-94

D slashed throat of elderly, rather feeble victim, almost decapitating him; also inflicted other knife wounds. D then broke into victim's house, threatened and tied up victim's wife, took money and victim's truck. D claimed self-defense.

1, 0, 2

.28

B27

.60

7. Fugate v. State,

431 S.E.2d 104

(Ga. 1993)

6-21-93

D broke into ex-wife's house, struck her at least 50 times, dragged her outside, grabbed her hair, pulled her head back, and shot her in the forehead. D claimed shooting was accidental.

1, 0, 2

.28

B4

.00

or

B12

.42

8. Osborne v. State,

430 S.E.2d 576

(Ga. 1993)

6-21-93

D shot his two companions in a car, one shot to each of their heads.

1, 0, 2

.28

C1

.32

9. Hill v. State,

427 S.E.2d 770

(Ga. 1993)

3-15-93

D bludgeoned cellmate with a nail-embedded board. D had earlier been convicted of anther capital offense.

2, 0, 2

.56

B15

(insufficient # of cases for valid proportion)

10. Davis v. State,

426 S.E.2d 844

(Ga. 1993)

2-26-93

D struck a man with a pistol and was chased by a police officer. When officer ordered halt, D shot officer, then, while smiling, approached officer and shot him several more times to eliminate him as a witness.

1, 1, 2

.81

B28

.69

11. Ward v. State,

417 S.E.2d 130

(Ga. 1992)

6-11-92

D kidnapped 5-month pregnant woman from her home, tied her up and transported her--she died from suffocation from wadded up paper towels stuffed down her throat. Months later, D kidnapped another woman from her home and raped her, boasting of having killed two people and dumping their bodies. When arrested, D had elaborate written plans for more attacks. D claimed lack of memory and remorse.

2, 1, 2

.86

B28

.69

12. Tharpe v. State,

416 S.E.2d 78

(Ga. 1992)

3-17-92

D threatened wife and she took out peace warrant. As she and her sister-in-law were driving, they encountered D in his vehicle. D blocked them, got out of car with shotgun, took sister-in-law from car and shot her three times, then drove wife away and raped her.

1, 0, 2

.28

B28

.69

13. Bennett v. State,

414 S.E.2d 218

(Ga. 1992)

3-13-92

D stabbed his sleeping wife over 100 times, then bludgeoned her with hammer. D claimed he was suffering from delusion that she and a third person were plotting to kill him and that she was acting in self-defense.

1, 0, 2

.28

B10

.03

14. Lynd v. State,

414 S.E.2d 5

(Ga. 1992)

2-27-92

D shot his wife; when she showed signs of life, shot her again and put her in car trunk. When she began thumping around, he opened trunk and shot her again. After burying body he drove to another state and killed another woman.

2, 0, 2

.56

B28

.69

15. Hall v. State,

415 S.E.2d 158

(Ga. 1991)

12-3-91

D's 10-year-old son annoyed him with a toy. D shot son with shotgun while the boy pleaded for his life. D expressed immediate remorse, but also boasted that he had taught the boy a lesson. D had history of terrorizing family. D had history of alcohol abuse, and had blood alcohol level of .32 around time of shooting.

1, 0, 1

.01

B9

.00

16. Todd v. State,

410 S.E.2d 725

(Ga. 1991)

11-27-91

D plotted with his girlfriend to tie D's roommate up and steal his car after the roommate fell asleep. Tired of waiting for roommate to fall asleep, D bludgeoned him in the head with a hammer at least 12 times, took victim's car and personal items, drove to another state where he hit another person in the head with a hammer, shot him, and took his wallet.

1, 0, 2

.28

B27

.60

17. Gibson v. State,

404 S.E.2d 781

(Ga. 1991)

6-10-91

D didn't like storekeeper victim and decided to rob and hurt him. D stabbed victim 39 times, including a wound in the neck that severed his spinal cord. D claimed he had used crack cocaine that day.

1, 1, 2

.81

B27

.60

18. Taylor v. State,

404 S.E.2d 255

(Ga. 1991)

5-10-91

D's wife obtained a "good-behavior warrant" against him. Immediately thereafter, he stabbed her numerous times, including a neck slash that almost decapitated her. D had a history of psychological and substance abuse problems.

0, 0, 2

.05

B10

.03

19. Ferrell v. State,

401 S.E.2d 741

(Ga. 1991)

3-15-91

D shot and killed his grandmother and cousin in order to rob the grandmother.

1, 0, 2

.28

C2

.88

20. Stripling v. State,

401 S.E.2d 500

(Ga. 1991)

2-22-91

D shot four of his co-workers, killing two of them while robbing the employer's place of business. He then stole a car at gunpoint. D claimed insanity and mental retardation.

0, 1, 2

.30

C2

.88

21. Pitts v. State,

386 S.E.2d 351

(Ga. 1989)

12-5-89

On same day he was released on parole, D kidnapped a woman from a parking lot, raped her, and killed her with a blow to the head with a pipe.

1, 1, 2

.81

B24

.25

22. Hightower v.

State, 386 S.E.2d

509 (Ga. 1989)

11-30-89

D shot his wife and two stepdaughters to death (ages of daughters not specified).

1, 0, 2

.28

C1

.32

23. Hall v. State,

383 S.E.2d 128

(Ga. 1989)

9-11-89

D vowed to kill his estranged wife, and boasted that he would "not get more than ten years." He broke into her house and stabbed her 17 times, including a series of wounds to her throat in a "necklace" pattern.

2, 0, 2

.56

B12

.42

24. Hatcher v. State,

379 S.E.2d 775

(Ga. 1989)

5-25-89

D shot and killed his former live-in lover and her mother in the mother's home. Defendant had been drinking.

1, 0, 2

.28

C1

.32

25. Jarrells v. State,

375 S.E.2d 842

(Ga. 1989)

2-8-89

D robbed two sisters who were his neighbors. He stabbed them both, tied them up, and beat them both on the head with a clothing iron. They were discovered the next day; one was dead.

1, 1, 2

.81

B28

.69

APPENDIX C[260]