The opinion of the court was delivered by: Poritz, C.j.
On proportionality review of a death sentence imposed in the Superior Court, Law Division, Middlesex County.
In State v. Chew, 150 N.J. 30 (1997) (Chew I), we affirmed John Chew's conviction and death sentence for the murder of Theresa Bowman. We "preserve[d] defendant's challenge to the proportionality of his death sentence" for later review, id. at 88; see N.J.S.A. 2C:11-3e, and now find no disproportionality.
The facts are set forth in detail in Chew I, supra, 150 N.J. at 42-50. We repeat here only that which is relevant to our proportionality review.
On the morning of January 13, 1993, the police found the body of Theresa Bowman in the driver's seat of John Chew's Corvette. The car was parked in the rear of the Woodbridge Hilton parking lot. Ms. Bowman's throat had been slashed approximately ten hours earlier. An eyewitness informed the police that he had seen a man "grabbing" a woman in the car. When the police interviewed Chew at his home that afternoon, he told them that he had last seen Ms. Bowman on the evening of January 12, 1993. Chew claimed that Ms. Bowman had accompanied him to the home of his sister, Crystal Charette, and then had driven off alone in his Corvette. Chew said that after Ms. Bowman left, he remained with his sister and her roommate, Helen Borden, for an hour-and-a-half and then both women drove him to his residence.
After the police obtained Chew's first statement, Chew returned to his sister's home. He told his sister and her roommate that he expected to be blamed for something he had not done. He said that a drug deal had gone bad and that he needed an alibi. He asked the two women to tell the police that they had been with him on the evening of January 12, and that he had remained at home the entire evening. They agreed.
On January 14, 1993, the police spoke to Chew's sister and her roommate. At this first meeting, the women corroborated Chew's version of his activities on the night of the murder.
Later that day, further questioning of Chew produced a taped statement. Chew repeated his initial story and provided information about his relationship with Ms. Bowman. At the time of the murder, the couple had been living together for over four years.
On January 15, 1993, the police received several telephone calls implicating Chew in Ms. Bowman's murder. The first call came from an insurance salesman who had sold the couple a joint life insurance policy in 1991. Under the policy, Chew was to receive $250,000 on the death of Ms. Bowman. The agent informed the police that on New Year's Eve 1992, just thirteen days before the murder, Chew had stopped at the agent's home and had asked to pay the December premium in cash. Chew told the agent that his check had bounced and that he did not want the policy to lapse. The agent told the police that no customer had ever come directly to his home with cash before.
George Tilton, a former employee of Chew's, also called the police. Tilton informed the police that in 1991, Chew had on numerous occasions offered him $10,000 to kill Ms. Bowman. According to Tilton, Chew wanted to kill his girlfriend in order to collect on her life insurance.
The third call came from Chew's son, Robert Chew, who at the time was incarcerated at the Ocean County Jail. Robert Chew said that in December 1991 his father told him about a plan to kill Ms. Bowman for life insurance proceeds.
The police received yet additional corroboration from Randy F., a Linden mechanic, who said that he and Ms. Bowman were having an affair. According to Randy F., Ms. Bowman had planned to leave Chew and move in with him as soon as Chew recovered a monetary settlement from an unrelated lawsuit. At trial, Randy F. testified that Ms. Bowman phoned him on January 12 and told him that she and Chew were driving to a location on the Garden State Parkway to pick up Chew's settlement check. All of this information suggested both a motive and a reason for the timing of Ms. Bowman's murder: Chew wanted the $250,000 insurance payment due on her death and feared that she would soon leave him and cancel the policy.
On January 23, the police arrested Chew, and a team of investigators again questioned his sister and her roommate. This time, the women told a different story, one that implicated John Chew. Chew's sister said that Chew called her from his home on the night of the murder and asked her to pick him up at the Woodbridge Hilton. Chew told his sister that he and Ms. Bowman were going to the hotel to pick up her paycheck and that he wanted to return immediately, but Bowman wished to remain with friends. When Chew's sister arrived with Borden at the hotel parking lot, she saw Chew get out of his Corvette. He was not injured, but there was blood on his clothing. Chew removed his outer clothes, put them into a plastic bag, and instructed Borden to put bleach in the bag. He then discarded the bag. Chew's sister told the police that after they returned to Chew's home, her brother told her what to tell the police and threatened her.
Borden's story was similar. She said she heard a "scream," that she believed came from the Corvette. About a minute later, Chew ran out of the Corvette and got into the car with his sister and Borden.
After being confronted with the inculpatory statements of his sister and her roommate, Chew agreed to give a taped statement, but refused to discuss the murder. His first account merely placed him at the scene of the crime and acknowledged that his sister and her roommate had driven him home from the Woodbridge Hilton. The trial court held that this statement was inadmissible because it was obtained in violation of defendant's right to counsel.
Later, after the taped statement concluded, Chew provided the police with a more detailed version of the events of January 12. This time Chew said that he and Ms. Bowman had gone to the Woodbridge Hilton on the night of the murder to complete a drug deal. Ms. Bowman was to handle the deal because only she knew the other drug dealer. Chew stated that he waited inside the Hilton, and that when he returned to the car Ms. Bowman was dead.
Chew unsuccessfully attempted to contact his lawyer and then began to talk about the crime again. He spoke of the cocaine deal, but this time acknowledged that Ms. Bowman was alive when he returned to the Corvette. According to Chew, Ms. Bowman claimed she was "ripped off" by the buyer and she and Chew began to quarrel. During this argument Ms. Bowman told Chew that she was having an affair with Randy F. Chew then "went off" on her.
After Chew gave this new account, the police readministered Miranda warnings and Chew signed a waiver. He gave another taped statement in which he again acknowledged driving to the Woodbridge Hilton with Ms. Bowman on the night of January 12 to sell a kilo of cocaine. As earlier, Chew told the police that he did not know the prospective buyer because the buyer was Ms. Bowman's contact. He said that Ms. Bowman waited in the car for the buyer and that he stayed in the doorway of the Hilton. When he returned to the Corvette after about forty-five minutes, Ms. Bowman told him that "she got ripped off," and they began to argue. Chew said that Ms. Bowman hit him "a couple of times in [the] face" and scratched him in the chin. He claimed he did not remember stabbing her.
After the fight, Chew left with his sister and her roommate who, at his request, had come to pick him up at the Hilton. He remembered being frightened because of all the blood, but denied knowing whether he had committed the murder. Chew also recalled getting rid of his bloody clothing and asking his sister and her roommate not to talk to the police about his whereabouts on the night of the murder. He claimed he did not remember threatening them, but acknowledged that "there's so much I don't remember." At the Miranda hearing, two corrections officials testified that Chew did not seem to be under the influence of drugs or undue stress when he made this statement; rather, he appeared calm and cooperative.
Defendant was indicted for purposeful or knowing murder by his own conduct, possession of a weapon for an unlawful purpose, terroristic threats, and other offenses later dismissed by the trial court. The prosecutor served notice of one aggravating factor: that defendant killed Ms. Bowman as consideration for the receipt, or in expectation of the receipt, of anything of pecuniary value. N.J.S.A. 2C:11-3c(4)(d). On June 13, 1995, the jury returned a verdict of guilty on two counts:
"purposeful or knowing murder by defendant's own conduct, in violation of N.J.S.A. 2C:11-3a(1) or (2), and possession of a weapon for an unlawful purpose, in violation of N.J.S.A. 2C:39-4d."
At the penalty-phase trial, the defense informed the jury, through a member of the parole board, that thirty years in prison was Chew's potential alternative sentence. A number of witnesses testified about the defendant's childhood and personal relationships.
"A forensic social worker interviewed family members and reviewed family records. She described a family history of beatings, drinking binges, abuse, unfaithfulness, and lack of love or encouragement. Defendant's mother and sister testified similarly."
"Defendant's eleven-year-old daughter Valerie spoke of her love for her father and of her contacts with him. A family therapist concluded that Valerie's continued contacts with her father would have a positive effect and that the two of them enjoyed a positive relationship. Defendant gave an allocution in which he sought mercy for the sake of his daughter." [Chew I, supra, 150 N.J. at 49-50.]
The jury found Chew guilty of murdering Ms. Bowman in expectation of the receipt of anything of pecuniary value. The jury also found ten mitigating factors under the statutory catchall factor, N.J.S.A. 2C:11-3c(5)(h). Having unanimously determined that the single aggravating factor outweighed the mitigating factor, the jury returned a death penalty verdict. The court merged the weapon conviction with the murder conviction and sentenced Chew to death. Defendant appealed directly to this Court as of right, R. 2:2-1(a)(3), and we affirmed his conviction and sentence, Chew I, supra, 150 N.J. at 88.
In Chew I we noted and preserved defendant's request for proportionality review, ibid., see N.J.S.A. 2C:11-3e. We now find that defendant has not shown his death sentence to be disproportionate.
The principal goal of proportionality review "is to determine whether a particular defendant's death sentence is disproportionate" when compared to the sentences of other defendants who are similarly situated. State v. DiFrisco, 142 N.J. 148, 160 (1995) (DiFrisco III); see N.J.S.A. 2C:11-3e. By this comparison, the Court seeks "'to ensure that the death penalty is being administered in a rational, non-arbitrary, and evenhanded manner, fairly and with reasonable consistency.'" State v. Loftin, 157 N.J. 253, 265 (1999) (Loftin II) (quoting State v. Marshall, 130 N.J. 109, 131 (1992) (Marshall II)). A defendant's death sentence is considered disproportionate if other defendants in the jurisdiction who have similar characteristics commit similar offenses and receive life sentences. State v. Martini, 139 N.J. 3, 20 (1994) (Martini II); State v. Bey, 137 N.J. 334, 343 (1994) (Bey IV); Marshall II, supra, 130 N.J. at 131. The defendant must show that his "death sentence is aberrational." Bey IV, supra, 137 N.J. at 352.
"There are two aspects of proportionality review: 'substantive,' or 'offense-oriented,' review; and 'procedural,' or 'offender-oriented,' review." DiFrisco III, supra, 142 N.J. at 160 (quoting Marshall II, supra, 130 N.J. at 126). Substantive or offense-oriented review examines "the offense to determine whether the punishment imposed is excessive in relation to the crime itself." Ibid.; see also Martini II, supra, 139 N.J. at 20. Procedural or offender-oriented review focuses on the defendant. "In such review, the question is 'whether the punishment fits the criminal.'" DiFrisco III, supra, 142 N.J. at 161 (quoting Marshall II, supra, 130 N.J. at 129) (additional internal quotations omitted). Our review here is offender oriented.
Our first step in any proportionality review is to determine the universe of cases that we will use to compare with the defendant's case. The 1992 amendment to N.J.S.A. 2C:11-3e limits this comparison group to only those cases in which a death sentence has actually been imposed. In Loftin II, however, we indicated that we were uncertain whether this limitation would "preclude meaningful appellate review." 157 N.J. at 264. We noted that our cases prior to Loftin II had "established the size of the proportionality review universe to include both death-eligible defendants and defendants who proceed to a penalty trial. . . . [and] that such broad categories would provide the most useful information about how decisions are made in the capital sentencing system by prosecutors and by juries." Id. at 264-65 (citations omitted). Because questions had been raised about the efficacy of the present system, we asked a special master to examine our proportionality review methodologies and report to the Court. Id. at 265. Until our review of the special master's findings and recommendations is complete, and we are able to determine the effect of the 1992 amendment on our review function, we "will continue to compare all death-eligible homicides with the case before us." Id. at 317.
The Administrative Office of the Courts ("AOC") compiled the data that we will use in this case in the Chew/Cooper/Harvey Report. The report includes all death-eligible cases known to the AOC as of July 31, 1997. Memorandum from Joseph J. Barraco, Esq., Acting Assistant Director, AOC Criminal Practice Division and Nina Rossi, Esq., Assistant Chief, Criminal Court Services, Criminal Practice Division, to Stephen W. Townsend, Clerk of the New Jersey Supreme Court 1 (Dec. 3, 1997) (Barraco Memorandum) (on file with the AOC). There were, as of that date, 401 death-eligible cases, of which 163 or forty-one percent, proceeded to a penalty trial. Chew/Cooper/Harvey Report, tbl. 3. Of the 163 penalty-trial cases, fifty, or thirty-one percent, resulted in a death sentence. Id., tbl. 2. The overall death-sentencing rate was therefore twelve percent (50/401). Id., tbl. 1.
B. Method of Classifying Cases
Once the universe of cases is determined, a database is developed from the facts of the comparison cases. In considering these data, "we use two approaches: an a priori approach and an empirical approach." Loftin II, supra, 157 N.J. at 323; see also DiFrisco III, supra, 142 N.J. at 163-64; Martini II, supra, 139 N.J. at 24; Bey IV, supra, 137 N.J. at 345; Marshall II, supra, 130 N.J. at 141-42. We have learned through experience the factors that affect capital-sentencing decisions, and in the a priori approach, we examine cases based on those factors. Loftin II, supra, 157 N.J. at 323; DiFrisco III, supra, 142 N.J. at 164; Martini II, supra, 139 N.J. at 24; Bey IV, supra, 137 N.J. at 345; Marshall II, supra, 130 N.J. at 141-42. "In the empirical method, we review both defendants who were sentenced to death and those who were not to 'identify those characteristics that determine the patterns'" behind disparate sentencing outcomes. Loftin II, supra, 157 N.J. at 323-24 (quoting DiFrisco III, supra, 142 N.J. at 164); see also Martini II, supra, 139 N.J. at 24; Bey IV, supra, 137 N.J. at 345; Marshall II, supra, 130 N.J. 142-43. By this method we sort the cases based on the factors found by prosecutors and juries to be "most relevant." Marshall II, supra, 130 N.J. at 143.
Consistent with our past practice, reversed death penalty cases "where the prosecutor chose not to proceed capitally on remand" are included in the death-sentenced universe. Loftin II, supra, 157 N.J. at 324; see also DiFrisco III, supra, 142 N.J. at 164; Martini II, supra, 139 N.J. at 25-26; Bey IV, supra, 137 N.J. at 345-49; Marshall II, supra, 130 N.J. at 194 n.10. We continue to believe that the original death penalty verdict is a useful reflection of jurors' views of a defendant's deathworthiness. Loftin II, supra, 157 N.J. at 324. Also, we will, as before, consider data that includes the defendant and data that excludes him. Ibid.
Having established both the universe of cases and the standards by which the cases are coded, we "'next group those cases according to their comparative levels of blameworthiness.'" Ibid. (quoting Martini II, supra, 139 N.J. at 28). In this analysis "[w]e measure blameworthiness according to statutory aggravating and mitigating factors as well as non-statutory factors based on objectively-verified measures of blameworthiness." Id. at 324-25 (internal quotations omitted). We evaluate these factors using two approaches: frequency analysis and precedent-seeking review. In frequency analysis, we compute the rate of death sentencing in cases similar to defendant's case in an attempt to "measure the societal consensus that death is the appropriate penalty in the measured cases." Bey IV, supra, 137 N.J. at 350. In precedent-seeking review, we engage in a traditional case-by-case analysis and "compare the defendant's case to factually-similar cases to determine whether the defendant is deathworthy in light of similarly-situated defendants." Martini II, supra, 139 N.J. at 28. As we discussed in Loftin II, supra, 157 N.J. at 296, because of concerns about the reliability of frequency review, we have focused on precedent-seeking review in deciding whether a defendant's death sentence is disproportionate.
A. Adjustments in Comparison Group
The AOC places the various death-eligible defendants in categories and subcategories based on the aggravating factors present in their cases. See Chew/Cooper/Harvey Report, tbl. 7. There are thirteen categories, each of which contains two to seven subcategories. *fn1
For the purposes of frequency review, defendant has not asked the Court to make any changes to the AOC category assignments. The State, however, objects to the AOC's inclusion of Walter Williams and Michael Rose in the pecuniary-motive category, and of James Clausell IB in the death-eligible universe. We agree with the State that Williams and Rose should be deleted from Chew's comparison group because in both cases the jury rejected the pecuniary-motive aggravating factor at defendants' penalty trials. See also DiFrisco III, supra, 142 N.J. at 170 (deleting Williams from DiFrisco's pecuniary-motive comparison group because "the jury rejected the pecuniary-motive aggravating factor at his penalty trial"). *fn2 Chew also agrees that Williams does not belong in his comparison group. As to the inclusion of Clausell 1B, because the defendant was retried for capital murder he was properly included in the death-eligible universe.
For the purposes of precedent-seeking review, defendant has suggested that we consider an additional thirty-four cases. Thirty-two of those cases are robbery murders, a type of murder that is significantly different from pecuniary-motive murders. See Chew I, supra, 150 N.J. at 56. Indeed, pecuniary-motive murders are three-and-a-half times more likely to result in death sentences than are robbery-murders. Chew/Cooper/Harvey Report, tbl. 7. As to the other two cases, one was a burglary murder in which the prosecution did not submit the pecuniary-motive aggravating factor to the jury, and the other involved multiple victims and resulted in an AOC placement in the A(1) multiple victim category. Again, we decline to alter the decision to place a defendant in only one comparison category. DiFrisco III, supra, 142 N.J. at 167. *fn3
We seek to determine through frequency analysis whether there is a societal consensus that the defendant in the case before us is sufficiently culpable such that his sentence may be deemed not aberrational. We therefore compare the defendant's culpability with that of other defendants, those who have been sentenced to death and those who have not. "The basic question in the frequency approach is whether the degree of blameworthiness in the instant case reasonably supports an expectation that such a case will generally result in a death sentence." Martini II, supra, 139 N.J. at 30. Some statistical disparity is permissible, indeed expected. "Proportionality review seeks to determine only whether a particular death sentence is aberrational, not whether it compares perfectly with other sentences." Bey IV, supra, 137 N.J. at 352. For this reason, "the Court has repeatedly declined to set a threshold at which the imposition of the death penalty becomes disproportionate." Loftin II, supra, 157 N.J. at 322; see also DiFrisco III, supra, 142 N.J. at 171-72; Martini II, supra, 139 N.J. at 30; Bey IV, supra, 137 N.J. at 351; Marshall II, supra, 130 N.J. at 152-54. Frequency analysis is not designed to provide "answers, but guidelines." DiFrisco III, supra, 142 N.J. at 172.
In Martini II, supra, 139 N.J. at 31-32, we explained more fully the statistical techniques used in frequency analysis. We will not repeat that explanation here except to note that before Loftin II, frequency review consisted of three tests: the salient-factors test; the numerical-preponderance test; and the index-of-outcomes test. See DiFrisco III, supra, 142 N.J. at 171; Martini II, supra, 139 N.J. at 29-30; Bey IV, supra, 137 N.J. at 350-51; Marshall II, supra, 130 N.J. at 154. In Loftin II, supra, 157 N.J. at 295, we eliminated the numerical-preponderance test because of its inherent limitations. We decided to conduct frequency review by means of the salient-factors and index-of-outcomes tests even as the Special Master we appointed developed recommendations in respect of these two tests. 157 N.J. at 317. Although the Special Master has submitted his report, Honorable David S. Baime, Report to the New Jersey Supreme Court: Proportionality Review Project (Apr. 28, 1999), we have not yet heard oral argument on the recommendations it contains, and therefore decline to comment on or use the report in this case. We note that both the defendant and the Dissent, post at ___ (slip op. at 26), call into question the validity of the frequency approach, pointing to small sample sizes and large confidence intervals. For a Discussion on confidence intervals see infra at ___ (slip op. at 32 n.6). We too remain concerned about the statistical reliability of frequency analysis, and shall, as in our prior proportionality review opinions, rely principally on precedent-seeking review. Loftin II, supra, 157 N.J. at 296; DiFrisco III, supra, 142 N.J. at 171; Martini II, supra, 139 N.J. at 28-29; Bey IV, supra, 137 N.J. at 350; Marshall II, supra, 130 N.J. at 173-74.
"In the salient-factors test, we compare defendant's sentence to sentences in factually-similar cases in order to measure the relative frequency of defendant's sentence." DiFrisco III, supra, 142 N.J. at 172. In this test, we first categorize cases based on statutory aggravating factors, and then further subdivide the group "'according to circumstances that serve either to aggravate or to mitigate the blameworthiness of the defendants in those cases.'" Loftin II, supra, 157 N.J. at 328 (quoting Martini II, supra, 139 N.J. at 33). The test is designed to determine the relative likelihood that a defendant committing a certain type of crime in a particular manner will receive the death penalty. Because the salient-factors test utilizes factually similar cases as a basis for comparison, we have "consistently viewed [this test] as the most persuasive of the frequency tests." Ibid.; see also DiFrisco III, supra, 142 N.J. at 173; Martini II, supra, 139 N.J. at 33; Bey IV, supra, 137 N.J. at 353; Marshall II, supra, 130 N.J. at 168.
Chew is categorized as a pecuniary-motive killer.
Chew/Copper/Harvey Report, tbl. 7A. The pecuniary-motive category is further divided into four subcategories: I-1, "a contract killing with defendant the killer;" I-2, "a contract killing with defendant the principal;" I-3, "defendant's motive was to obtain a pecuniary advantage (e.g., inheritance) occurring as a matter of law upon the victim's death;" and I-4, "the victim paid the defendant to kill him or her." Id., tbl. 6. The AOC has classified Chew as an I-3 killer. Id., tbl. 7A. Excluding Walter Williams from the I-3 subcategory, see supra at ___ (slip op. at 18), only one other defendant, Celestine Payne, remains. For this reason, any statistical analysis that considered only the I-3 subcategory would be unproductive. Accordingly, we will compare defendant with the entire group of pecuniary-motive killers. See, e.g., DiFrisco III, supra, 142 N.J. at 174 (comparing DiFrisco, a pecuniary-motive category, contract-killer subcategory, murderer to the entire category of pecuniary-motive murderers). *fn4
Of the sixteen eligible I category cases, nine proceeded to the penalty-trial stage, and of the nine penalty-trial cases, five resulted in death sentences, including defendant's case. The death-sentencing rate for all death-eligible pecuniary-motive killers is thirty-one percent, and the death-sentencing rate for those advancing to the penalty-trial stage is fifty-five percent. These rates are considerably higher than the overall death-sentencing rates of twelve percent for all death-eligible killers and thirty-one percent for all penalty-phase cases. Removing defendant's case from the group lowers the rates somewhat, but when we look at the overall death-sentencing rates without defendant there is no discernable difference.
In tabular form, the figures are:
Death-Sentencing Rates for All Pecuniary-Motive Killers
(Data adjusted* from Chew/Cooper/Harvey Report, tbl. 7)
Death-Sentencing Rate Death-Sentencing Rate Proportion of Cases at Penalty Trial for All Eligible Cases Advancing to P-Trial
I Incl. D .55 (5/9) .31 (5/16) .56 (9/16)
I Excl. D .50 (4/8) .27 (4/15) .53 (8/15)
All Ds .31 (50/163) .12 (50/401) .41 (163/401)
All Ds but D .30 (49/162) .12 (49/400) .41 (162/400)
*The data has been adjusted to reflect the exclusion of Williams and Rose. See supra at __ (slip op. at 18).
We observe that the death penalty is imposed more frequently in the pecuniary-motive category than in the general death-eligible and penalty-phase groups, and that juries regularly find pecuniary motive killers to be deathworthy.
"While we are mindful that the small sample sizes prevent us from relying on the results of this test," Loftin II, supra, 157 N.J. 329, because a significant proportion of defendants in the pecuniary-motive category have received the death penalty, we conclude that the test does not demonstrate disproportionality.
2. Index-of-Outcomes Test
The index-of-outcomes test "'compare[s] cases that are factually dissimilar but that are nevertheless comparable from the perspective of the defendants' blameworthiness.'" Id. at 330 (quoting Martini II, supra, 139 N.J. at 42). "Unlike the other tests, the basis for comparison here is not similar factual patterns or numerical indices but 'a roughly-equivalent measure of blameworthiness.'" Martini II, supra, 139 N.J. at 42 (quoting Marshall II, supra, 130 N.J. at 172). "We compare the blameworthiness of different defendants by the statistically-relevant measures of culpability found in the circumstances of their cases." Ibid.; see also Bey IV, supra, 137 N.J. at 362.
By this approach, we attempt to weigh statutory and non-statutory, aggravating and mitigating "factors according to the influence each carries with prosecutors and jurors." DiFrisco III, supra, 142 N.J. at 178.
In conducting the index-of-outcomes test, we can use two different sets of data for the culpability estimates of prior defendants. In Martini II, supra, 139 N.J. at 45, we utilized the data compiled in each prior defendant's AOC report for the culpability estimates of that defendant. Since Martini II, however, we have utilized the data compiled by the AOC in its most recent report. See Loftin II, supra, 157 N.J. at 331-34; DiFrisco III, supra, 142 N.J. at 180-82. Although an individual defendant's culpability measures may vary slightly from report to report, utilizing the data in the Chew/Cooper/Harvey Report will enable us to assess Chew's culpability relative to other death-sentenced defendants based on the most recent estimates. See Loftin II, supra, 157 N.J. at 289 (stating "[a]s a matter of general principle, the broadest possible statistical database should provide the most useful information"). While the variations in the reports over this relatively short period of time do not alter the results of our analyses, they do contribute to our growing concern regarding the usefulness of the index-of-outcomes test. Nevertheless, as indicated supra at ___ (slip op. at 21-22), we will continue to employ the test in its present form until we complete our review of the Special Master's findings and recommendations.
The Chew/Cooper/Harvey Report contains tables prepared by the AOC that are based on multivariate models *fn5 that measure the defendants' culpability. Chew/Cooper/Harvey Report, technical app. 9 at 1. The cases are placed into five levels of culpability, based on the predicted probability for each defendant that a jury would return a sentence of death. Those culpability levels are: level one, less than a twenty percent likelihood of a death sentence being returned; level two, twenty to less-than-forty percent; level three, forty to less-than-sixty percent; level four, sixty to less-than-eighty percent; and, level five, eighty to one hundred percent. Cases involving similar levels of blameworthiness are thereby grouped together for comparison purposes. DiFrisco III, supra, 142 N.J. at 179; Martini II, supra, 139 N.J. at 42-43.
As in our prior proportionality review cases, the AOC urges caution in relying on the results of these analyses, stating that "the addition of cases over time has had a positive impact on the stability of the models, [but] the culpability estimate which purports to give a 'predicted probability of a death sentence' is often still too soft, and little substantive reliance should be given to this statistic in the Chew, Cooper, and Harvey cases." Barraco Memorandum, supra, at 4; see also Loftin II, supra, 157 N.J. at 330. We therefore once again must "'treat the index-of-outcomes findings accordingly.'" DiFrisco III, supra, 142 N.J. at 179 (quoting Martini II, supra, 139 N.J. at 43).
The results of the index-of-outcomes test in tabular form are as follows:
INDEX-OF-OUTCOME TEST RESULTS
(data from Chew/Cooper/Harvey Report, tbls. 13-14, 21-25)
Statutory & Non-statutory Factors Statutory Factors Only
Penalty Trial Death-Eligible Penalty Trial Death-Eligible
Predicted Probability .67 .19 .90 .47
Range (.22 to .94) (.02 to .72) (.51 to .99) (.09 to .88)
Culpability Level 4 1 5 3
Death-Sentencing .75 (9/12) .05 (18/334) .76 (13/17) .52 (9/17)
Predicted Probability .21 .44 .72 .53
Range (.01 to .92) (.12 to .83) (.26 to .95) (.17 to .85)
Culpability Level 2 3 4 3
Death-Sentencing .26 (6/23) .55 (6/11) .63 (12/19) .52 (9/17)
Predicted Probability .68 .19 .46 .19
Range (.24 to .93) (.02 to .78) (.08 to .89) (.01 to .84)
Culpability Level 4 1 3 1
Death-Sentencing .75 (9/12) .05 (18/334) .45 (13/29) .05 (15/318)
Predicted Probability .81 .09 .26 .12
Range (.36 to .97) (.02 to .32) (.09 to .55) (.04 to .31)
Culpability Level 5 1 2 1
Death-Sentencing .77 (23/30) .05 (18/334) .31 (10/32) .05 (15/318)
Predicted Probability .65 .38 .74 .40
Range (.22 to .93) (.12 to .74) (.40 to .92) (.18 to .68)
Culpability Level 4 2 4 3
Death-Sentencing .75 (9/12) .33 (10/30) .63 (12/19) .52 (9/17)
Predicted Probability .64 .13 .59 .27
Range (.15 to .95) (.00 to .87) (.12 to .94) (.03 to .82)
Culpability Level 4 1 3 2
Death-Sentencing .75 (9/12) .05 (18/334) .45 (13/29) .26 (12/45)
Considering both statutory and non-statutory factors in penalty-trial cases, defendant has a sixty-seven percent predicted probability of receiving a death sentence, with a probability range of twenty-two percent to ninety-four percent. Chew/Cooper/Harvey Report, tbl. 22. The predicted probability of sixty-seven percent places Chew in culpability level four. Ibid. The penalty-trial death sentencing rate at that culpability level is seventy-five percent (9/12). Id., tbl. 21.
Utilizing the same factors and universe, a comparison between Chew and our prior proportionality review defendants reveals that Chew's rates are comparable to those of DiFrisco, Bey and Marshall, somewhat lower than those of Martini and far higher than those of Loftin. Id., tbl. 22. These figures do not demonstrate that Chew's sentence is disproportionate.
Considering again both statutory and non-statutory factors, but expanding the universe to all death-eligible cases, Chew has a death sentence predicted probability of nineteen percent, with a probability range of two percent to seventy-two percent. Id., tbl. 14. That places defendant in culpability level one, ibid., which has a death-sentencing rate of five percent (18/334), id., tbl. 13. Although only five percent of defendants in culpability level one have received the death penalty, it is important to remember that the overall death-sentencing rate is just twelve percent (50/401). Ibid. Further, comparing Chew to previous proportionality review defendants shows that the predicted probability of a death sentence and the death-sentencing rates for Loftin and Bey are somewhat higher than defendant's, and that defendant's rates exceed those of Martini and Marshall and are substantially identical to those of DiFrisco. Id., tbl. 14. This comparison also does not support a finding of disproportionality.
Considering only the statutory aggravating and mitigating factors, and limiting the universe to penalty-trial cases, defendant has a predicted probability of death of ninety percent, with a probability range of fifty-one percent to ninety-nine percent. Id., tbl. 24. That places defendant in culpability level five, ibid., which has a death-sentencing rate of seventy-six percent (13/17), id., tbl. 23. Considering the statutory factors only, and using the same penalty-trial universe, Chew's probabilities exceed those of the other proportionality review defendants and do not support a finding of disproportionality. Id., tbl. 24.
Finally, we look at defendant's case examining only the statutory factors, but using the entire death-eligible universe. Chew has a predicted probability of death of forty-seven percent, with a probability range of nine percent to eighty-eight percent. Id., tbl. 25. That places defendant in culpability level three, which has a death-sentencing rate of fifty-two percent (9/17). Id. at tbls. 23, 25. Chew's figures are substantially identical to those of Loftin, and exceed those of the other defendants. Id., tbl. 25. Once again, we do not find disproportionality.
We are satisfied that the index-of-outcomes test does not indicate that Chew's sentence of death is aberrational. Nevertheless, we remain wary of this test. As in the previous proportionality review cases, "'the small sample size of cases with similar levels of blameworthiness and the great ranges in the confidence intervals' *fn6 prevent us from relying on the results." Loftin II, supra, 157 N.J. at 295-96 (quoting DiFrisco III, supra, 142 N.J. at 182). Moreover, the variability of defendant's predicted probability of a death sentence further suggests that the ...