Opinion of Souter, J. STRICKLER v. GREENE ____ U. S. ____ (1999) SUPREME COURT OF THE UNITED STATES No. 98–5864

TOMMY DAVID STRICKLER, PETITIONER v.FRED W. GREENE, WARDEN

on writ of certiorari to the united states court of appeals for the fourth circuit

[June 17, 1999]

Justice Souter, with whom Justice Kennedy joins as to Part II, concurring in part and dissenting in part.

I look at this case much as the Court does, starting with its view in Part III (which I join) that Strickler has shown cause to excuse the procedural default of his Brady claim. Like the Court, I think it clear that the materials withheld were exculpatory as devastating ammunition for impeaching Stoltzfus. See ante, at 19. Even on the question of prejudice or materiality, over which I ultimately part company with the majority, I am persuaded that Strickler has failed to establish a reasonable probability that, had the materials withheld been disclosed, he would not have been found guilty of capital murder. See ante, at 29–32. As the Court says, however, the prejudice enquiry does not stop at the conviction but goes to each step of the sentencing process: the jury’s consideration of aggravating, death-qualifying facts, the jury’s discretionary recommendation of a death sentence if it finds the requisite aggravating factors, and the judge’s discretionary decision to follow the jury’s recommendation. See ante, at 31–33. It is with respect to the penultimate step in determining the sentence that I think Strickler has carried his burden. I believe there is a reasonable probability (which I take to mean a significant possibility) that disclosure of the Stoltzfus materials would have led the jury to recommend life, not death, and I respectfully dissent.

I

Before I get to the analysis of prejudice I should say something about the standard for identifying it, and about the unfortunate phrasing of the shorthand version in which the standard is customarily couched. The Court speaks in terms of the familiar, and perhaps familiarly deceptive, formulation: whether there is a “reasonable probability” of a different outcome if the evidence withheld had been disclosed. The Court rightly cautions that the standard intended by these words does not require defendants to show that a different outcome would have been more likely than not with the suppressed evidence, let alone that without the materials withheld the evidence would have been insufficient to support the result reached. See ante, at 27; Kyles v. Whitley, 514 U. S. 419, 434–435 (1995). Instead, the Court restates the question (as I have done elsewhere) as whether “ ‘the favorable evidence could reasonably be taken to put the whole case in such a different light as to undermine confidence’ ” in the outcome. Ante, at 27 (quoting Kyles, supra, at 435).

Despite our repeated explanation of the shorthand formulation in these words, the continued use of the term “probability” raises an unjustifiable risk of misleading courts into treating it as akin to the more demanding standard, “more likely than not.” While any short phrases for what the cases are getting at will be “inevitably imprecise,” United States v. Agurs, 427 U. S. 97, 108 (1976), I think “significant possibility” would do better at capturing the degree to which the undisclosed evidence would place the actual result in question, sufficient to warrant overturning a conviction or sentence.

To see that this is so, we need to recall Brady’s evolution since the appearance of the rule as originally stated, that “suppression by the prosecution of evidence favorable to an accused upon request violates due process where the evidence is material either to guilt or to punishment, irrespective of the good faith or bad faith of the prosecution.” Brady v. Maryland, 373 U. S. 83, 87 (1963). Brady itself did not explain what it meant by “material” (perhaps assuming the term would be given its usual meaning in the law of evidence, see United States v. Bagley, 473 U. S. 667, 703, n. 5 (1985) (Marshall, J., dissenting)). We first essayed a partial definition in United States v. Agurs, supra, where we identified three situations arguably within the ambit of Brady and said that in the first, involving knowing use of perjured testimony, reversal was required if there was “any reasonable likelihood” that the false testimony had affected the verdict. Agurs, supra, at 103 (citing Giglio v. United States, 405 U. S. 150, 154 (1972), in turn quoting Napue v. Illinois, 360 U. S. 264, 271 (1959)). We have treated “reasonable likelihood” as synonymous with “reasonable possibility” and thus have equated materiality in the perjured-testimony cases with a showing that suppression of the evidence was not harmless beyond a reasonable doubt. Bagley, supra, at 678–680, and n. 9 (opinion of Blackmun, J.). See also Brecht v. Abrahamson, 507 U. S. 619, 637 (1993) (defining harmless-beyond-a-reasonable-doubt standard as no “ ‘reasonable possibility’ that trial error contributed to the verdict”); Chapman v. California, 386 U. S. 18, 24 (1967) (same). In Agurs, we thought a less demanding standard appropriate when the prosecution fails to turn over materials in the absence of a specific request. Although we refrained from attaching a label to that standard, we explained it as falling between the more-likely-than-not level and yet another criterion, whether the reviewing court’s “ ‘conviction [was] sure that the error did not influence the jury, or had but very slight effect.’ ” 427 U. S., at 112 (quoting Kotteakos v. United States, 328 U. S. 750, 764 (1946)). Finally, in United States v. Bagley, supra, we embraced “reasonable probability” as the appropriate standard to judge the materiality of information withheld by the prosecution whether or not the defense had asked first. Bagley took that phrase from Strickland v. Washington, 466 U. S. 668, 694 (1984), where it had been used for the level of prejudice needed to make out a claim of constitutionally ineffective assistance of counsel. Strickland in turn cited two cases for its formulation, Agurs (which did not contain the expression “reasonable probability”) and United States v. Valenzuela-Bernal, 458 U. S. 858, 873–874 (1982) (which held that sanctions against the Government for deportation of a potential defense witness were appropriate only if there was a “reasonable likelihood” that the lost testimony “could have affected the judgment of the trier of fact”).

The circuitous path by which the Court came to adopt “reasonable probability” of a different result as the rule of Brady materiality suggests several things. First, while “reasonable possibility” or “reasonable likelihood,” the Kotteakos standard, and “reasonable probability” express distinct levels of confidence concerning the hypothetical effects of errors on decisionmakers’ reasoning, the differences among the standards are slight. Second, the gap between all three of those formulations and “more likely than not” is greater than any differences among them. Third, because of that larger gap, it is misleading in Brady cases to use the term “probability,” which is naturally read as the cognate of “probably” and thus confused with “more likely than not,” see Morris v. Mathews, 475 U. S. 237, 247 (1986) (apparently treating “reasonable probability” as synonymous with “probably”); id., at 254, n. 3 (Blackmun, J., concurring in judgment) (cautioning against confusing “reasonable probability” with more likely than not). We would be better off speaking of a “significant possibility” of a different result to characterize the Brady materiality standard. Even then, given the soft edges of all these phrases, the touchstone of the enquiry must remain whether the evidentiary suppression “undermines our confidence” that the factfinder would have reached the same result.

II

Even keeping in mind these caveats about the appropriate level of materiality, applying the standard to the facts of this case does not give the Court easy answers, as the Court candidly acknowledges. See ante, at 26. Indeed, the Court concedes that discrediting Stoltzfus’s testimony “might have changed the outcome of the trial,” ante, at 27, and that the District Court was “surely correct” to find a “reasonable possibility that either a total, or just a substantial, discount of Stoltzfus’ testimony might have produced a different result, either at the guilt or sentencing phases,” ante, at 28–29.

In the end, however, the Court finds the undisclosed evidence inadequate to undermine confidence in the jury’s sentencing recommendation, whereas I find it sufficient to do that. Since we apply the same standard to the same record, our differing conclusions largely reflect different assessments of the significance the jurors probably ascribed to the Stoltzfus testimony. My assessment turns on two points. First, I believe that in making the ultimate judgment about what should be done to one of several participants in a crime this appalling the jurors would very likely have given weight to the degree of initiative and leadership exercised by that particular defendant. Second, I believe that no other testimony comes close to the prominence and force of Stoltzfus’s account in showing Strickler as the unquestionably dominant member of the trio involved in Whitlock’s abduction and the aggressive and moving figure behind her murder.

Although Stoltzfus was not the prosecution’s first witness, she was the first to describe Strickler in any detail, thus providing the frame for the remainder of the story the prosecution presented to the jury. From the start of Stoltzfus’s testimony, Strickler was “Mountain Man” and his male companion “Shy Guy,” labels whose repetition more than a dozen times (by the prosecutor as well as by Stoltzfus) must have left the jurors with a clear sense of the relative roles that Strickler and Henderson played in the crimes that followed Stoltzfus’s observation. According to her, when she first saw Strickler she “just sort of instinctively backed up because I was frightened.” App. 36. Unlike retiring “Shy Guy,” Strickler was “revved up.” Id., at 39, 60. Even in describing her first encounter with Strickler inside the mall, Stoltzfus spoke of him as domineering, a “very impatient” character yelling at his female companion, “Blonde Girl,” to join him. Id., at 36, 38–39.

After describing in detail how “Mountain Man” and “Blonde Girl” were dressed, Stoltzfus said that “ ‘Mountain Man’ came tearing out of the Mall entrance door and went up to the driver of [a] van and . . . was just really mad and ran back and banged on back of the backside of the van” while “Shy Guy” and “Blonde Girl” hung back. Id., at 43. “Mountain Man” approached a pickup truck, then “pounded on” the front passenger side window of Whitlock’s car, “shook and shook the car door,” “banging and banging on the window” while Whitlock checked to see if the door was locked. Ibid. Finally, “he just really shook it hard and you could tell he was mad. Shook it really hard and the door opened and he jumped in . . . and faced her.” Id., at 43–44. While Whitlock tried to push him away, “Mountain Man” “motioned for ‘Blonde Girl’ and ‘Shy Guy’ to come” and the girl did as she was bidden. She “started to jump into the car,” but “jumped back” when Whitlock stepped on the gas. Id., at 44. Then “Mountain Man” started “hitting [Whitlock] on the left shoulder, her right shoulder and then . . . the head,” finally “open[ing] the door again” so “the ‘Blonde Girl’ got in the back and ‘Shy Guy’ followed and got behind him.” Id., at 45. “Shy Guy” passed “Mountain Man” his tan coat, which “Mountain Man” “fiddled with” for “what seemed like a long time,” then “sat back up and . . . faced” Whitlock while “the other two in the back seat sat back and relaxed.” Ibid. Stoltzfus then claimed that she got out of her car and went over to Whitlock’s, whereupon unassertive “Shy Guy” “instinctively jumped, you know, laid over on the seat to hide from me.” Id., at 46. Stoltzfus pulled up next to Whitlock’s car and repeatedly asked, “[A]re you O.K.[?],” but Whitlock responded only with eye contact; “she didn’t smile, there was no expression,” and “[j]ust very serious, looked down to her right,” suggesting Strickler was holding a weapon on her. Id., at 46, 47. Finally, Whitlock mouthed something, which Stoltzfus demonstrated for the jury and then explained she realized must have been the word, “help.” Id., at 47.

Without rejecting the very notion that jurors with discretion in sentencing would be influenced by the relative dominance of one accomplice among others in a shocking crime, I could not regard Stoltzfus’s colorful testimony as anything but significant on the matter of sentence. It was Stoltzfus alone who described Strickler as the initiator of the abduction, as the one who broke into Whitlock’s car, who beckoned his companions to follow him, and who violently subdued the victim while “Shy Guy” sat in the back seat. The bare content of this testimony, important enough, was enhanced by one of the inherent hallmarks of reliability, as Stoltzfus confidently recalled detail after detail. The withheld documents would have shown, however, that many of the details Stoltzfus confidently mentioned on the stand (such as Strickler’s appearance, Whitlock’s appearance, the hour of day when the episode occurred, and her daughter’s alleged notation of the license plate number of Whitlock’s car) had apparently escaped her memory in her initial interviews with the police. Her persuasive account did not come, indeed, until after her recollection had been aided by further conversations with the police and with the victim’s boyfriend. I therefore have to assess the likely havoc that an informed cross-examiner could have wreaked upon Stoltzfus as adequate to raise a significant possibility of a different recommendation, as sufficient to undermine confidence that the death recommendation would have been the choice. All it would have taken, after all, was one juror to hold out against death to preclude the recommendation actually given.

The Court does not, of course, deny that evidence of dominant role would probably have been considered by the jury; the Court, instead, doubts that this consideration, and the evidence bearing on it, would have figured so prominently in a juror’s mind as to be a fulcrum of confidence. I am not convinced by the Court’s reasons.

The Court emphasizes the brutal manner of the killing and Strickler’s want of remorse, as jury considerations diminishing the relative importance of Strickler’s position as ringleader. See ante, at 33. Without doubt the jurors considered these to be important factors, and without doubt they may have been treated as sufficient to warrant death. But as the Court says, sufficiency of other evidence and the facts it supports is not the Brady standard, and the significance of both brutality and sangfroid must surely have been complemented by a certainty that without Strickler there would have been no abduction and no ensuing murder.

The Court concludes that Stoltzfus’s testimony is unlikely to have had significant influence on the jury’s sentencing recommendation because the prosecutor made no mention of her testimony in his closing statement at the sentencing proceeding. See ante, at 33. But although the Court is entirely right that the prosecution gave no prominence to the Stoltzfus testimony at the sentencing stage, the State’s closing actually did include two brief references to Strickler’s behavior in “just grabbing a complete stranger and abducting her,” 19 Record 919; see also id., at 904, as relevant to the jury’s determination of future dangerousness. And since Strickler’s criminal record had no convictions involving actual violence, a point defense counsel stressed in his closing argument, see id., at 913, the jurors may well have given weight to Stoltzfus’s lively portrait of Strickler as the aggressive leader of the group, when they came to assess his future dangerousness.

What is more important, common experience, supported by at least one empirical study, see Bowers, Sandys, & Steiner, Foreclosed Impartiality in Capital Sentencing: Jurors’ Predispositions, Guilt-Trial Experience, and Premature Decision Making, 83 Cornell L. Rev. 1476, 1486–1496 (1998), tells us that the evidence and arguments presented during the guilt phase of a capital trial will often have a significant effect on the jurors’ choice of sentence. True, Stoltzfus’s testimony directly discussed only the circumstances of Whitlock’s abduction, but its impact on the jury was almost certainly broader, as the prosecutor recognized. After the jury rendered its verdict on guilt, for example, the defense moved for a judgment of acquittal on the capital murder charge based on insufficiency of the evidence. In the prosecutor’s argument to the court he replied that

“the evidence clearly shows that this man was the aggressor. He was the one that ran out. He was the one that grabbed Leanne Whitlock. When she struggled trying to get away from him … , he was the one that started beating her there in the car. And finally subdued her enough to make her drive away from the mall, so you start with the principle that he is the aggressor.” 20 Record 15.

Stoltzfus’s testimony helped establish the “principle,” as the prosecutor put it, that Strickler was “the aggressor,” the dominant figure, in the whole sequence of criminal events, including the murder, not just in the abduction. If the defense could have called Stoltzfus’s credibility into question, the jurors’ belief that Strickler was the chief aggressor might have been undermined to the point that at least one of them would have hesitated to recommend death.

The Court suggests that the jury might have concluded that Strickler was the leader based on three other pieces of evidence: Kurt Massie’s identification of Strickler as the driver of Whitlock’s car on its way toward the field where she was killed; Donna Tudor’s testimony that Strickler kept the car the following week; and Tudor’s testimony that Strickler threatened Henderson with a knife later on the evening of the murder. But if we are going to look at other testimony we cannot stop here. The accuracy of both Massie’s and Tudor’s testimony was open to question, and all of it was subject to some evidence that Henderson had taken a major role in the murder. The Court has quoted the District Court’s summation of evidence against him, ante, at 28, n. 36: Henderson’s wallet was found near the body, his clothes were bloody, he presented a woman friend with the victim’s watch at a postmortem celebration (which he left driving the victim’s car), and he confessed to a friend that he had just killed an unidentified black person. Had this been the totality of the evidence, the jurors could well have had little certainty about who had been in charge. But they could have had no doubt about the leader if they believed Stoltzfus.

Ultimately, I cannot accept the Court’s discount of Stoltzfus in the Brady sentencing calculus for the reason I have repeatedly emphasized, the undeniable narrative force of what she said. Against this, it does not matter so much that other witnesses could have placed Strickler at the shopping mall on the afternoon of the murder, ante, at 31, or that the Stoltzfus testimony did not directly address the aggravating factors found, ante, at 33. What is important is that her evidence presented a gripping story, see E. Loftus & J. Doyle, Eyewitness Testimony: Civil and Criminal 5 (3d ed. 1997) (“[R]esearch redoundingly proves that the story format is a powerful key to juror decision making”). Its message was that Strickler was the madly energetic leader of two morally apathetic accomplices, who were passive but for his direction. One cannot be reasonably confident that not a single juror would have had a different perspective after an impeachment that would have destroyed the credibility of that story. I would accordingly vacate the sentence and remand for reconsideration, and to that extent I respectfully dissent.

 The Court notes that the District Court did not resolve whether all eight of the Stoltzfus documents had been withheld, as Strickler claimed, or only five. For purposes of its decision granting summary judgment for Strickler, the District Court assumed that only five had not been disclosed. See ante, at 27, 15. The Court of Appeals also left the dispute unresolved, see App. 418, n. 8, though granting summary judgment for respondent based on a lack of prejudice would presumably have required that court to assume that all eight documents had been withheld. Because this Court affirms the grant of summary judgment for respondent based on lack of prejudice and because it relies on at least one of the disputed documents in its analysis, see ante, at 19, I understand it to have assumed that none of the eight documents was disclosed. I proceed based on that assumption as well. If one thought the difference between five and eight documents withheld would affect the determination of prejudice, a remand to resolve that factual question would be necessary.

 In keeping with suggestions in a number of our opinions, see Schlup v. Delo, 513 U. S. 298, 327, n. 45 (1995); Sawyer v. Whitley, 505 U. S. 333, 345 (1992), the Court treats the prejudice enquiry as synonymous with the materiality determination under Brady v. Maryland, 373 U. S. 83 (1963). See ante, at 19, 26–27, 34. I follow the Court’s lead.

 Each of these phrases or standards has been used in a number of contexts. This Court has used “reasonable possibility,” for example, in defining the level of threat of injury to competition needed to make out a claim under the Robinson-Patman Act, see, e.g., Brooke Group Ltd. v. Brown & Williamson Tobacco Corp., 509 U. S. 209, 222 (1993); the standard for judging whether a grand jury subpoena should be quashed under Federal Rule of Criminal Procedure 17(c), see United States v. R. Enterprises, Inc., 498 U. S. 292, 301 (1991); and the debtor’s burden in establishing that certain collateral is necessary to reorganization and thus exempt from the Bankruptcy Code’s automatic stay provision, see United Sav. Assn. of Tex v. Timbers of Inwood Forest Associates, Ltd., 484 U. S. 365, 375–376 (1988). We have adopted the standard established in Kotteakos v. United States, 328 U. S. 750 (1946), for determining the harmlessness of nonconstitutional errors on direct review as the criterion for the harmlessness enquiry concerning constitutional errors on collateral review. See Brecht v. Abrahamson, 507 U. S. 619, 637–638 (1993). We have used “reasonable probability” to define the plaintiff’s burden in making out a claim under §7 of the Clayton Act, see, e.g., Brown Shoe Co. v. United States, 370 U. S. 294, 325 (1962); FTC v. Morton Salt Co., 334 U. S. 37, 55–61 (1948) (Jackson, J., dissenting in part) (contrasting “reasonable possibility” and “reasonable probability” and arguing for latter as appropriate standard under Robinson-Patman Act); the standard for granting certiorari, vacating, and remanding in light of intervening developments, see, e.g., Lawrence v. Chater, 516 U. S. 163, 167 (1996) (per curiam); and the standard for exempting organizations from otherwise valid disclosure requirements in light of threats or harassment resulting from the disclosure, see, e.g., Buckley v. Valeo, 424 U. S. 1, 74 (1976) (per curiam). We have recently used “significant possibility” in explaining the circumstances under which nominal compensation is an appropriate award in a suit under the Longshore and Harbor Workers’ Compensation Act, see Metropolitan Stevedore Co. v. Rambo, 521 U. S. 121, 123 (1997), but we most commonly use that term in defining one of the requirements for the granting of a stay pending certiorari. The three-part test requires a “reasonable probability” that the Court will grant certiorari or note probable jurisdiction, a “significant possibility” that the Court will reverse the decision below, and a likelihood of irreparable injury absent a stay. See, e.g., Barefoot v. Estelle, 463 U. S. 880, 895 (1983); Packwood v. Senate Select Comm. on Ethics, 510 U. S. 1319 (1994) (Rehnquist, C. J., in chambers).

 Massie’s identification was open to some doubt because it occurred at night as one car passed another on a highway. Moreover, he testified that he first saw four people in the car, then only three, and that none of the occupants was black. App. 66–67, 70–73. Tudor, as defense counsel brought out on cross-examination, testified pursuant to a cooperation agreement with the government and admitted that the story she told on the stand was different from what she had told the defense investigator before trial. Id., at 100–101, 103–104.