Tillman v. State

128 P.3d 1123 (2005) 2005 UT 56

ElRoy TILLMAN, Plaintiff and Appellee,
v.
STATE of Utah, Defendant and Appellant.

No. 20030148.

Supreme Court of Utah.

August 30, 2005. Rehearing Denied November 29, 2005.

Loni F. DeLand, Michael R. Sikora, McCaye Christianson, Salt Lake City, for plaintiff.

Mark Shurtleff, Att'y Gen., Erin Riley, Thomas Brunker, Asst. Att'ys Gen., Salt Lake City, for defendant.

DURRANT, Justice:

¶ 1 In 1983, following a jury trial, ElRoy Tillman was convicted of capital murder and sentenced to death. After he exhausted all avenues of relief under state and federal law, Tillman was scheduled to be executed. However, shortly before his execution date, Tillman discovered that the State had failed to disclose transcripts of two interviews it had conducted with the key prosecution witness prior to trial. Citing those transcripts, Tillman filed a petition for post-conviction relief with the district court. Although the court rejected Tillman's assertion that the undisclosed evidence was sufficient to warrant a reversal of his conviction, the court nevertheless agreed that he was entitled to relief with respect to his sentence of death. As a result, the district court vacated Tillman's sentence and ordered a new sentencing proceeding.

¶ 2 The State challenges the district court's ruling, arguing that Tillman's petition is procedurally barred under Utah's Post-Conviction Remedies Act or, alternatively, that the suppression of the transcripts did not violate Tillman's due process rights. Because we agree with the district court that Tillman's petition is not procedurally barred and that the suppression of the transcripts violated his due process rights, we affirm.

BACKGROUND

¶ 3 On May 26, 1982, Mark Schoenfeld was found dead in his Salt Lake City apartment. Police arrested Tillman for the murder based on information provided by Lori Groneman, a former girlfriend of Tillman who was dating Schoenfeld at the time of his death. Police also arrested Carla Sagers, Tillman's girlfriend at the time of the murder.

¶ 4 Sagers initially confirmed the alibi Tillman had given to the police. However, in exchange for full immunity, Sagers recanted that confirmation. Sagers became the State's key witness at trial, testifying that on the night of May 25, 1982, she and Tillman entered Schoenfeld's home, where they found Schoenfeld asleep in his bed. According to Sagers's testimony, Tillman bludgeoned Schoenfeld twice in the head with an ax and then set the bed on fire while Schoenfeld was still alive. Due to the lack of forensic evidence linking Tillman to the crime scene, Sagers's testimony was unquestionably the most critical evidence the State presented at trial. In fact, as noted by Justice Stewart, "[t]he only direct evidence of Tillman's involvement in the crime, indeed the only evidence of his involvement at all came from the testimony of Carla Sagers." Tillman v. Cook, 855 P.2d 211, 228 (Stewart, J., dissenting). The jury no doubt relied heavily upon Sagers's testimony in finding Tillman guilty of first-degree murder. Because the jury returned a guilty verdict, the trial court initiated a separate sentencing proceeding, at which the trial jury considered aggravating and mitigating circumstances bearing on the appropriateness of the death penalty. See Utah Code Ann. § 76-3-207 (Supp.1982) (current version at Utah Code Ann. § 76-3-207 (2003)). After considering all the evidence, the jury unanimously agreed to impose the death penalty. The trial judge subsequently sentenced Tillman to death.

¶ 5 After Tillman exhausted all avenues of state and federal relief in the years following his conviction,^[1] his execution date was set for June 24, 2001. On May 5, 2001, Tillman filed a petition for commutation before the Utah Board of Pardons and Parole. In connection with that petition, Tillman requested reports that the State had created when it administered polygraph examinations to Sagers prior to trial. In response, the State provided defense counsel with two uncertified, undated, typed partial transcripts of pre- or post-polygraph interviews that Sergeant Kenneth Thirsk had conducted with Sagers in early December 1982 and January 1983, shortly before Tillman's trial began on January 4, 1983.^[2] Although the transcripts were found in one of the prosecution's case files, neither transcript had been previously disclosed to defense counsel.^[3] In fact, at trial, the State introduced testimony that no recordings of the interviews were ever made.

¶ 6 Following the disclosure of the transcripts, the parties stipulated to a stay of Tillman's execution date and Tillman filed a petition for post-conviction relief with the district court. In his petition, Tillman argued that the State's failure to disclose the partial transcripts violated his right to due process under Brady v. Maryland, 373 U.S. 83, 83 S. Ct. 1194, 10 L. Ed. 2d 215 (1963), and its progeny, which recognize that prosecutors have a constitutional duty to disclose all material evidence that is favorable to a defendant. Because of that violation, Tillman argued that he was entitled to a new trial, or alternatively, to a reduced sentence of life imprisonment.^[4]

¶ 7 The district court agreed with Tillman in part. In an articulate and well-reasoned opinion, the district court began by rejecting the contention that Tillman's claim was procedurally barred by Utah's Post-Conviction Remedies Act, which provides that "[a] person is not eligible for relief ... upon any ground that ... could have been, but was not, raised in a previous request for post-conviction relief." Utah Code Ann. § 78-35a-106(d) (1999). Although the State asserted that defense counsel could have discovered the undisclosed partial transcripts by subpoenaing the prosecutor's files in preparation for Tillman's prior appeals and post-conviction petitions, the district court reasoned that "[i]t belie[d] common sense to suggest that the defense could have discovered th[e] evidence when the polygrapher and the State affirmatively denied the existence of the same." Consequently, the court concluded that Tillman had demonstrated by a preponderance of the evidence that his claim could not have been previously raised and that it was, therefore, not procedurally barred.

¶ 8 The district court further ruled that the State's failure to disclose the transcripts violated Tillman's rights under the Due Process Clause of the Fourteenth Amendment. Applying the Brady test, the court first reasoned that the suppressed evidence was favorable to the defense because the transcripts contained previously unknown information that could have been used to impeach Sagers's testimony at trial. Although the district court acknowledged that the information contained in the partial transcripts was, "for the most part, the same information that was used during trial to attack [Sagers's] truthfulness," the court noted that the transcripts contained other non-cumulative impeachment evidence favorable to Tillman. Specifically, the court observed that (1) Sgt. Thirsk displayed a great deal more disbelief and incredulity at Sagers's account of the murder in the transcripts than he displayed at trial, (2) some of Sgt. Thirsk's statements and questions in the transcripts gave the appearance that he was coaching Sagers into giving more believable testimony, and (3) the partial transcripts contained over sixty notations indicating that Sagers had laughed while being questioned about the murder. The district court concluded that the impeachment value of this information was favorable to Tillman and therefore satisfied the first component of the Brady test.

¶ 9 The court also determined that the second component of the Brady test was satisfied because the undisclosed evidence was suppressed by the State. Although it was unclear who made the recordings and had them transcribed, the court concluded that, under the circumstances, the interviews must have been recorded by someone affiliated with either law enforcement or the prosecution. As a result, the district court reasoned that the prosecutors either knew or should have known of the existence of the transcripts, and that they therefore either knowingly or inadvertently suppressed the evidence by failing to disclose it prior to trial.

¶ 10 Addressing Brady's final component, the district court concluded that the State's failure to disclose the partial transcripts prior to trial prejudiced Tillman. The court was "unconvinced that a different outcome of the guilt phase of the trial would have resulted even if the additional impeachment evidence had been known and utilized by defense counsel." Nevertheless, the court was persuaded that the probability of prejudice resulting from the undisclosed partial transcripts was sufficiently high to undermine the court's confidence in the death sentence imposed during the penalty phase of Tillman's trial.

¶ 11 In reaching this conclusion, the court focused on what it identified as the inherent relationship between Sagers's credibility and her moral culpability for Schoenfeld's murder. The court reasoned that "the less credible [Sagers's] testimony was shown to be, the more likely her degree of moral culpability for the homicide would have risen in the minds of the jurors," and that "the greater her degree of moral culpability, the less likely jurors would have voted to impose a sentence of death upon [Tillman]." The court noted that, in order to impose the death penalty, jurors must unanimously agree that the totality of aggravating evidence outweighs, beyond a reasonable doubt, the totality of mitigating evidence. The court believed that the mitigating evidence calling Sagers's moral culpability into question probably would have altered, in Tillman's favor, at least one juror's assessment of the aggravating and mitigating factors. Therefore, the court concluded there was a reasonable probability that disclosure of the partial transcripts would have resulted in a different outcome during the penalty phase of Tillman's trial.

¶ 12 After determining that Tillman had met his burden by establishing all three Brady components, the district court ruled that the State's failure to disclose the partial transcripts violated Tillman's due process rights. Consequently, the court vacated Tillman's sentence and ordered a new sentencing proceeding. The State filed a timely notice of appeal.

¶ 13 On appeal, the State argues that the district court erred in vacating Tillman's sentence for two reasons. First, it argues that Tillman has failed to prove that he could not have raised his current Brady claim in either of his two previous state post-conviction petitions for relief. As a result, the State contends that Tillman's current petition for post-conviction relief is procedurally barred. Second, and alternatively, the State argues that its failure to disclose the partial transcripts does not constitute a Brady violation because Tillman cannot establish that he was prejudiced by the nondisclosure. We have jurisdiction to review the district court's decision pursuant to Utah Code section 78-2-2(3)(i) (2002).

STANDARD OF REVIEW

¶ 14 On appeal from a ruling on a petition for post-conviction relief, we review the post-conviction court's legal conclusions for correctness and its factual findings for clear error. Julian v. State, 2002 UT 61, ¶ 8, 52 P.3d 1168.

ANALYSIS^[5]

I. TILLMAN'S POST-CONVICTION PETITION IS NOT PROCEDURALLY BARRED BY UTAH'S POST-CONVICTION REMEDIES ACT

¶ 15 Before addressing whether the State's failure to disclose the partial transcripts violated Tillman's due process rights, we first turn to the State's contention that Tillman should be procedurally barred from asserting his Brady claim in this petition for post-conviction relief.

¶ 16 Utah's Post-Conviction Remedies Act^[6] ("PCRA") provides, in relevant part, that a person who has been convicted and sentenced for a criminal offense may file a post-conviction petition requesting that the court modify or vacate either the conviction or sentence on the ground that "the conviction was obtained or the sentence was imposed in violation of the United States Constitution or Utah Constitution." Utah Code Ann. § 78-35a-104(1)(a) (2002). However, no relief may be granted upon any ground that "was raised or addressed in any previous request for post-conviction relief or could have been, but was not, raised in a previous request for post-conviction relief." Id. § 78-35a-106(1)(d) (emphasis added). The State has the burden of pleading preclusion. Id. § 78-35a-105. Once it has done so, the petitioner must disprove preclusion by a preponderance of the evidence. Id.

¶ 17 In this case, the State has raised a procedural bar defense; therefore, Tillman has the burden to prove that his current Brady claim could not have been raised in either of his previous state petitions for post-conviction relief. The State argues that Tillman has failed to meet this burden. It asserts that, prior to filing Tillman's two previous state petitions for post-conviction relief, reasonably diligent post-conviction counsel would have reviewed the prosecution's open files to determine if there were any issues that trial or appellate counsel had overlooked. According to the State, such a review would have uncovered the partial transcripts and enabled Tillman to bring a Brady claim in an earlier post-conviction petition. Because Tillman's post-conviction counsel admittedly did not review the files following Tillman's trial, the State argues that Tillman cannot meet his burden in establishing that his Brady claim survives the procedural bar.

¶ 18 Tillman counters that due diligence does not require a defendant to continually reexamine prosecution files for previously hidden or undisclosed exculpatory or impeachment evidence. Tillman states that this is especially true in this case, where the State explicitly represented at trial that no recordings of the pre- or post-polygraph interviews, the transcripts of which form the basis of his current claim, were made. Given this representation and the State's ongoing duty to disclose favorable material evidence, Tillman argues that he located the transcripts as soon as he reasonably could have done so.

¶ 19 In support of their respective arguments, both parties rely on federal case law relating to federal post-conviction preclusion. However, we need not delve into federal jurisprudence, as we can decide the issue based on existing Utah law.

¶ 20 Like the PCRA, Utah common law prohibits a petitioner from raising a post-conviction claim relating to the denial of a constitutional right when the claim could have been raised in a prior post-conviction proceeding. See Hurst v. Cook, 777 P.2d 1029, 1033 (Utah 1989) (citing Utah R. Civ. P. 65B(i)(4)). However, we have consistently recognized exceptions to this general rule in "unusual circumstances" where "good cause" excuses a petitioner's failure to raise the claim earlier. See id. at 1037. Currently, rule 65C of the Utah Rules of Civil Procedure maintains this exception, providing that "[a]dditional claims relating to the legality of [a] conviction or sentence may not be raised in subsequent [post-conviction] proceedings except for good cause shown." Utah R. Civ. P. 65C(c).

¶ 21 When evaluating a post-conviction claim for good cause, courts should generally decline to review a contention of error where the error "is something which is known or should have been known to the party," and therefore could have been raised at an earlier time. Brown v. Turner, 21 Utah 2d 96, 440 P.2d 968, 969 (1968). "Nevertheless, howsoever desirable it may be to adhere to the rules, the law should not be so blind and unreasoning that where an injustice has resulted the victim should be without remedy." Martinez v. Smith, 602 P.2d 700, 702 (Utah 1979); see also Hurst, 777 P.2d at 1036 ("[I]t has long been our law[] that a procedural default is not always determinative of a collateral attack on a conviction where it is alleged that the trial was not conducted within the bounds of basic fairness or in harmony with constitutional standards."). Thus, even where an issue could have been raised in a previous post-conviction petition, post-conviction review may be available in those "rare cases," Martinez, 602 P.2d at 702, or "unusual circumstances" where "an obvious injustice or a substantial and prejudicial denial of a constitutional right has occurred" that would make it "unconscionable" not to reexamine the issue, Hurst, 777 P.2d at 1035; cf. Andrews v. Shulsen, 773 P.2d 832, 833 (Utah 1988) (declining to review a petition for post-conviction relief where the petitioner failed to show good cause for not raising the claim of constitutional error in a previous post-conviction petition).

¶ 22 In Hurst, we identified five good cause factors sufficient to justify the filing of a successive claim not raised in a prior post-conviction petition. Those factors include

(1) the denial of a constitutional right pursuant to new law that is, or might be, retroactive[;] (2) new facts not previously known which would show the denial of a constitutional right or might change the outcome of the trial[;] (3) the existence of fundamental unfairness in a conviction[;] (4) the illegality of a sentence[;] or (5) a claim overlooked in good faith with no intent to delay or abuse the writ.

777 P.2d at 1037 (citations and footnote omitted). We later clarified, in Candelario v. Cook, 789 P.2d 710, 712 (Utah 1990), that Hurst's list of "good cause" factors is not exhaustive. Nevertheless, the PCRA has codified only the first two of the five "good cause" factors identified in Hurst. Gardner v. Galetka, 2004 UT 42, ¶ 14, 94 P.3d 263. However, because "the power to review post-conviction petitions ``quintessentially ... belongs to the judicial branch of government,'" id. at ¶ 17 (quoting Hurst, 777 P.2d at 1033), and not the legislature, all five common law exceptions "retain their independent constitutional significance and may be examined by this court in our review of post-conviction petitions," id. at ¶ 15.

¶ 23 In this case, we conclude that Tillman has demonstrated "good cause," warranting review of his current post-conviction petition on the merits. For even if we were to agree with the State that Tillman technically could have filed his current Brady claims in one of his previous petitions for post-conviction relief had his appellate counsel examined the prosecution's files, we remain unconvinced that, under the circumstances, Tillman should have done so.

¶ 24 The State admittedly failed to disclose to defense counsel prior to trial either the partial transcripts or the audio recordings from which those transcripts were made. The State compounded this failure when it allowed Sgt. Thirsk to testify during trial, without contradiction, that his conversations with Sagers were not recorded. The prosecutor himself later stated that he was unaware that recordings of the interviews had been made. Consequently, Tillman had no reason to believe that there were undisclosed transcripts until the State revealed their existence some nineteen years later—after the audio recordings and, possibly, the complete interview transcripts were lost or destroyed.

¶ 25 Under these circumstances, we are persuaded that Tillman's Brady claim was overlooked in good faith with no intent to delay or abuse the post-conviction process. We therefore hold that Tillman has demonstrated sufficient "good cause" to justify examining his petition for post-conviction relief on the merits.

¶ 26 Having determined that the claims raised in Tillman's post-conviction petition are not procedurally barred, we turn our analysis to whether the State's failure to disclose the partial transcripts violated Tillman's rights under the Due Process Clause of the Fourteenth Amendment.

II. THE STATE'S FAILURE TO DISCLOSE THE PARTIAL TRANSCRIPTS VIOLATED TILLMAN'S RIGHT TO DUE PROCESS

¶ 27 In Brady v. Maryland, 373 U.S. 83, 83 S. Ct. 1194, 10 L. Ed. 2d 215 (1963), the United States Supreme Court held that "the suppression by the prosecution of evidence favorable to an accused ... 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." Id. at 87, 83 S. Ct. 1194. The Court later clarified that the duty to disclose favorable evidence encompasses both exculpatory and impeachment evidence. United States v. Bagley, 473 U.S. 667, 676, 105 S. Ct. 3375, 87 L. Ed. 2d 481 (1985). The duty to disclose favorable evidence is implicated even if the evidence is known only to police investigators and not the prosecutor, Kyles v. Whitley, 514 U.S. 419, 438, 115 S. Ct. 1555, 131 L. Ed. 2d 490 (1995), and regardless of whether the evidence has been requested by the accused, United States v. Agurs, 427 U.S. 97, 107, 96 S. Ct. 2392, 49 L. Ed. 2d 342 (1976).

¶ 28 In Strickler v. Greene, 527 U.S. 263, 119 S. Ct. 1936, 144 L. Ed. 2d 286 (1999), the Court identified the following three components of a Brady prosecutorial misconduct claim: (1) the evidence at issue is "favorable to the accused, either because it is exculpatory, or because it is impeaching"; (2) the evidence was "suppressed by the State, either willfully or inadvertently"; and (3) prejudice ensued. Id. at 281-82, 119 S. Ct. 1936. On appeal, the State concedes that the first and second Brady requirements have been satisfied in this case. Consequently, the only remaining issue is whether Tillman suffered prejudice as a result of the State's failure to disclose the partial transcripts.

¶ 29 For evidence to be prejudicial for Brady purposes, it must be material. See id. at 282-83, 119 S. Ct. 1936. Evidence is material if "``there is a reasonable probability that, had the evidence been disclosed to the defense, the result of the proceeding would have been different.'" Kyles, 514 U.S. at 433, 115 S. Ct. 1555 (quoting Bagley, 473 U.S. at 682, 105 S. Ct. 3375). A reasonable probability of a different result occurs "when the government's evidentiary suppression ``undermines confidence in the outcome of the trial.'" Id. at 434, 115 S. Ct. 1555 (quoting Bagley, 473 U.S. at 678, 105 S. Ct. 3375).^[7]

¶ 30 In weighing whether evidence is "material" for Brady purposes, three principles deserve special emphasis. First, when determining whether a "reasonable probability" of a different result exists, "the question is not whether the defendant would more likely than not have received a different [result] with the evidence," but rather, "whether in its absence [the defendant] received a fair trial, understood as a trial resulting in a verdict worthy of confidence." Id.

¶ 31 Second, "materiality ... is not a sufficiency of the evidence test," id., and, therefore, "not just a matter of determining whether, after discounting the inculpatory evidence in light of the undisclosed evidence, the remaining evidence is sufficient to support the jury's conclusions," Strickler, 527 U.S. at 290, 119 S. Ct. 1936. To establish materiality, a defendant need only show "that the favorable evidence could reasonably be taken to put the whole case in such a different light as to undermine confidence in the verdict" or sentence. Kyles, 514 U.S. at 435, 115 S. Ct. 1555.

¶ 32 Third, the materiality of suppressed evidence must be evaluated in the context of the entire record. Agurs, 427 U.S. at 112, 96 S. Ct. 2392. Although a court may "evaluate the tendency and force of the undisclosed evidence item by item," it is the cumulative or collective effect of the evidence that is weighed when determining whether the disclosure would have created a reasonable probability of a different result. Kyles, 514 U.S. at 436 & n. 10, 115 S. Ct. 1555.

¶ 33 Applying the aforementioned principles to this case, the district court concluded that the undisclosed partial transcripts were "material," as Brady contemplates that term, to the sentencing phase of Tillman's trial. In reaching this conclusion, the district court determined that the transcripts provided non-cumulative impeachment evidence in the following three ways: (1) the transcripts suggested that Sgt. Thirsk disbelieved Sagers's account of the murder to a much greater degree than he testified to at trial; (2) the transcripts gave the appearance that Sgt. Thirsk had coached Sagers into giving more believable testimony; and (3) the transcripts contained numerous notations of laughter on the part of Sagers, suggesting inappropriate levity. The district court reasoned that this evidence would have undermined Sagers's credibility, thereby elevating her own degree of moral culpability for the murder and correspondingly diminishing the likelihood that jurors would have voted to impose the death penalty on Tillman.

¶ 34 The State challenges the district court's ruling on several grounds. First, it argues that the evidence in the transcripts reflecting Sgt. Thirsk's disbelief and the appearance of coaching is cumulative to other evidence available at trial, and that, even if the evidence was not cumulative, it would ultimately have been damaging to Tillman had it been presented at trial. Second, the State argues that the notations of Sagers's laughter in the transcripts likely reflected a nervous reaction to being interrogated, and that even if the transcripts reflected inappropriate levity, such levity would have had no correlation to Sagers's credibility. Finally, the State argues that none of the evidence contained in the partial transcripts impacts Sagers's moral culpability and that, even if it did, her moral culpability would have been irrelevant during the sentencing proceedings. As a result, the State contends that the absence of the undisclosed transcripts at trial does not undermine confidence in Tillman's death sentence.

¶ 35 Although we address each of these arguments below, we begin our analysis by first examining Tillman's assertion on appeal that the undisclosed transcripts contain additional non-cumulative impeachment evidence that the district court failed to identify. We then return to an examination of the district court's ruling and the alleged errors the State identifies therein.

A. Additional Evidence in the Undisclosed Partial Transcripts that Tillman Claims Is Non-Cumulative

¶ 36 In its opinion, the district court concluded that, aside from the three areas expressly identified in its ruling, the information contained in the partial transcripts relevant to Sagers's credibility was cumulative to information used to attack her truthfulness during trial. Tillman challenges this conclusion. He asserts that the suppressed transcripts contain admissions and statements that were not made at trial or in any of the evidence provided to the defense before trial,^[8] and that the information contained in the suppressed transcripts would have proved valuable in his attempt to impeach Sagers.

¶ 37 We agree with the district court that many of the statements Tillman identifies as non-cumulative are, in fact, cumulative of other impeachment evidence available at trial and, thus, do not constitute material evidence for Brady purposes. Nevertheless, as we explain below, we agree with Tillman that the transcripts provide additional important non-cumulative evidence in two respects. We first address the evidence that the district court correctly concluded was cumulative and then address the additional evidence that we deem non-cumulative.

1. Cumulative Evidence

¶ 38 According to Tillman, the suppressed transcripts contain evidence incorrectly classified by the district court as cumulative. The evidence identified by Tillman relates to four categories: (1) Sagers's confusion as to her location at the time Schoenfeld was struck, (2) the increased specificity in Sagers's trial testimony when compared with her testimony in the suppressed transcripts, (3) Sagers's role in the decision to set Schoenfeld's bed on fire, and (4) the absence of descriptions of Tillman as "revengeful" in the suppressed transcripts. We address each of these categories in turn.

a. Sagers's location at the time Schoenfeld was struck

¶ 39 Tillman first focuses on Sagers's admission in the transcripts that she may have been present in the bedroom when one of blows to Schoenfeld was struck. Specifically, in response to Sgt. Thirsk's suggestion that she was inside the room when Tillman hit Schoenfeld, Sagers responded that "I was on my way out. I turned around, I was on my way out.... I could of [sic] been in and I could of [sic] been out." Tillman argues that this statement could have been used to rebut Sagers's testimony in previous statements and at trial wherein she denied being in the room when the blows were struck.

¶ 40 Although we acknowledge that this statement could have been used to impeach Sagers's account of the murder at trial, we agree with the State that it was merely cumulative of other evidence presented at trial. As previously noted, the failure to disclose evidence does not, in all instances, result in a constitutional violation. See Bagley, 473 U.S. at 678, 105 S. Ct. 3375; see also United States v. Quintanilla, 193 F.3d 1139, 1149 (10th Cir.1999) (concluding that no Brady violation occurred where a defendant already knew of evidence contained in an undisclosed report prior to trial); United States v. Davis, 787 F.2d 1501, 1505 (11th Cir.1986) (concluding that no Brady violation occurred where the prosecution failed to disclose discrepancies between the grand jury and trial testimony of various government witnesses because the witnesses were exhaustively cross-examined as to those discrepancies at trial).

¶ 41 In this case, Tillman elicited testimony from Sgt. Thirsk at trial that Sagers had both admitted to and denied being in the bedroom at the time the blows were struck:

Q. Was there any question and any answer from [Sagers] in connection with whether or not she was in the room at the time of Mark Schoenfeld's death?

A. Yes, her response to me was that she had been in the room, in the bedroom but as to where it was at the exact time of death, there was no conversation regarding the time of death, but at the time blows were struck, that was the essence of the conversation.

Q. Had she indicated to you previously that she had not been in the room at the time the blows were struck?

A. Yes, she had told me previously she had not.

Because disclosure of the transcripts would not have augmented this testimony in any meaningful way, we conclude that the evidence contained in the suppressed transcripts illustrating Sagers's confusion as to her presence in Schoenfeld's room at the time the blows were struck is cumulative and not material for purposes of Brady.

b. Increased specificity in Sagers's trial testimony

¶ 42 Tillman next points to what he identifies as a differing level of specificity between Sagers's account in the suppressed transcripts and her testimony at trial when explaining how she assisted Tillman in covering Schoenfeld's head during the murder. In particular, he argues that Sagers stated in the transcripts that during the murder, Tillman asked for a "thing" or "something" for an unspecified reason, but at trial, she testified clearly that Tillman asked her to hand him an article of clothing lying on the floor, which he placed over Schoenfeld's head, presumably to prevent blood from splattering on the walls when he struck the second blow. Tillman argues that if the partial transcripts had been disclosed prior to trial, he could have used the information contained therein to point out the "sudden improvement" in Sagers's memory in the short time before the trial.^[9]

¶ 43 The problem with Tillman's argument is that the suppressed transcripts do not evidence a discrepancy in the level of specificity between Sagers's statements in the partial transcript and her trial testimony about Tillman's request for an item to place over Schoenfeld's head. Although Tillman is correct in his observation that Sagers stated in the transcripts only that Tillman asked her to hand him a "thing" or "something," both her preliminary hearing and trial testimony contain similarly vague descriptions.

¶ 44 For example, during the preliminary hearing, Sagers stated twice that Tillman requested she hand him "a shirt or something" that was lying on the floor. Additionally, she testified that when she and Tillman later burned a towel, which had been used to wipe down the walls of Schoenfeld's bedroom, they also burned the "shirt" or "whatever it was" that had been placed over Schoenfeld's head. Perhaps most importantly, at trial, Sagers testified that Tillman asked her "to hand him something ... something laying on the floor, a piece of clothing." In this context, the fact that in the suppressed transcripts Sagers described handing Tillman a "thing" or "something" would have provided no additional impeachment evidence at trial and therefore the district court correctly concluded that such evidence was cumulative and non-material for Brady purposes.

c. Lack of testimony referring to Tillman as "revengeful"

¶ 45 Tillman next argues that the partial transcripts provide valuable non-cumulative evidence because of what they do not contain. At trial, Sagers explained her involvement in the murder by testifying that Tillman was "revengeful" and by expressing her concern that Tillman might "do something to [her]" if she went to the police or otherwise attempted to thwart the murder plot. Tillman asserts that in the suppressed transcripts, "as in her four other transcribed pre-trial statements, Sagers does not explain her complicity in the murder as resulting from her knowledge of Tillman's ``revengefulness.'" (Emphasis added.) He argues that the "revengefulness" aspect of Sagers's testimony offered critical support to the prosecution's argument for the death penalty because it allowed the State to portray Tillman as a threat not just to Lori Groneman, Tillman's ex-girlfriend, but to the community at large.

¶ 46 Although we do not rely on three of the four transcribed statements identified by Tillman because they are not contained in the record, see supra ¶ 36 n. 8, we accept Tillman's concession that, like the suppressed transcripts, the other pretrial statements made no mention of Tillman's "revengefulness." We also observe that, during the preliminary hearing, Sagers never suggested Tillman was dangerous, but rather, testified that Tillman non-threateningly "told me that it was all right" after she refused to kill Schoenfeld and Groneman with a gun he provided.^[10] Because the transcripts possessed by Tillman prior to trial could have been used to demonstrate that Sagers's expressions of fear of Tillman's "revengefulness" at trial was a recent addition to her testimony. Thus, the omission of any reference to Tillman's "revengefulness" in the suppressed transcripts is merely cumulative of other evidence that defense counsel possessed and could have utilized during trial.^[11]

d. The decision to set Schoenfeld's bed on fire

¶ 47 Tillman asserts that another non-cumulative feature of the suppressed transcripts "pertains to Sagers'[s] role in the decision to set [Schoenfeld's] bed on fire." He argues that Sagers admitted at trial, for the first time, that the idea of using cigarettes and setting a fire was hers, a statement contradicting the partial transcripts, in which she insisted that she had "no involvement" in the homicide. Tillman contends that, if the suppressed transcripts had been disclosed prior to trial, defense counsel could have pointed to Sagers's ever-evolving story and argued that she "made the admission about the fire in response to extreme pressure" applied by Sgt. Thirsk during the pretrial interviews and "out of fear of violating [her] immunity agreement."

¶ 48 As an initial response, we reject Tillman's characterization of Sagers's trial testimony as an admission that the idea of setting a fire was her own. Although Sagers admitted during trial that it was her idea to stage the scene to make it appear that cigarettes had caused the fire, she never admitted to having suggested starting the fire itself. Additionally, Tillman admits that the absence of an admission regarding the cigarette idea is "consistent with the other recorded pre-trial statements" available to defense prior to trial. Because, by Tillman's own admission, these other statements could have been used in the place of the suppressed transcripts to challenge Sagers's "evolving story," we conclude that Sagers's failure to mention in the suppressed transcripts that the use of cigarettes was her idea is merely cumulative to impeachment evidence available to the defense at trial.

¶ 49 Having concluded that the categories of evidence discussed above constitute mere cumulative impeachment evidence, we now turn to the portions of the suppressed transcripts that we conclude provide non-cumulative impeachment evidence.

2. Non-Cumulative Evidence

¶ 50 Tillman argues that the undisclosed partial transcripts provide important non-cumulative impeachment evidence not identified by the district court. Specifically, Tillman argues that the suppressed transcripts (1) belie the reliability of Sagers's testimony recounting the order of events during the murder and (2) reveal that Tillman was suicidal at the time the murder was committed. We address each of Tillman's arguments in turn.

a. The sequence of events

¶ 51 Tillman first observes that the suppressed transcripts establish that Sagers was unclear about the sequence of events surrounding Schoenfeld's murder. According to Tillman, this uncertainty had all but vanished when Sagers took the stand at trial, at which time she was able to confidently recall the precise order of events. Tillman argues that the suppressed transcripts could have been used to show that Sagers's memory dramatically improved in a very short time, thereby undermining her credibility. We agree.

¶ 52 At trial, Sagers definitively testified that after hearing Tillman hit Schoenfeld the first time, she looked inside the bedroom and observed Tillman wiping blood off the wall with a towel. She stated that Tillman then asked her to hand him something lying on the floor—a piece of clothing—which he placed over Schoenfeld's head before striking him again with the ax.

¶ 53 We acknowledge that defense counsel could have utilized Sagers's preliminary hearing testimony to challenge, to some degree, the accuracy of Sagers's trial testimony addressing the order of events. At the preliminary hearing, Sagers testified that, after hearing Tillman hit Schoenfeld a "couple of times," she was asked to hand him "a shirt or something that was lying on the floor" before Tillman got a towel to wipe blood off the wall. However, as is evident from the following excerpt, Sagers displayed significant uncertainty as to the order of events in her pretrial interviews with Sgt. Thirsk:

Q. How did you know what to hand him?

A. (Sign) [sic]

Q. Did he say, I'm going to hit him again and you said put something over him I didn't want [sic] to see him?

A. No I didn't say that.

Q. O.k.

Q. But he said hand me what?

A. ... Now I don't know ... I don't know if he wiped the wall off with a towel before or after. Now I don't know.

Q. But you remember that for sure.

A. I remember him doing. [sic]

Q. Wiping the wall ...

A. But I don't remember if that was ... before he put this other thing on his head or after. I don't remember which ... how it came.

Q. You mean what? How does this refer in the order of when you handed it to him?

A. Well I don't know ...

(Ellipses in original.)

¶ 54 Unlike Sagers's statement during the preliminary hearing, which was made months before trial, this testimony was elicited, at most, four weeks prior to, and possibly mere days before, Sagers testified with seeming confidence at trial. We agree with Tillman that, given the statement's proximity to trial, it is not merely cumulative impeachment evidence, and we conclude that it could have been used to undermine Sagers's credibility at trial. We leave for later analysis whether, when considered with all other non-cumulative evidence contained in the undisclosed partial transcripts, this evidence is sufficient to undermine our confidence in the sentence Tillman received.

b. Sagers's statement that Tillman was suicidal

¶ 55 Tillman next argues that Sagers's statement in the suppressed transcripts that Tillman was suicidal in the weeks leading up to the murder is also important non-cumulative evidence. We agree.

¶ 56 In the suppressed transcripts, Sagers stated that Tillman would "get real depressed sometimes and well he did'nt [sic] really come out and say this one night but he, he said it's either got to be him or me ya know he was incinuating [sic] suicide the way I took it." Because the State does not challenge Tillman's assertion that this suicide reference is non-cumulative, and because we have not been directed to other evidence available prior to trial indicating that Tillman may have been suicidal in the weeks preceding the murder, we agree that this evidence was non-cumulative and could have been presented as a mitigating factor during the penalty phase of the proceedings. As with the evidence of Sagers's uncertainty about the sequence of events surrounding the murder, we leave for later analysis whether, in conjunction with all the additional non-cumulative, suppressed evidence, Sagers's statement that Tillman was suicidal prior to Schoenfeld's murder is sufficient to undermine our confidence in his sentence.

¶ 57 Having addressed the evidence Tillman asserts the district court incorrectly considered cumulative, we now examine whether the district court correctly identified three additional categories of non-cumulative evidence contained in the suppressed transcripts.

B. Evidence Identified as Non-Cumulative by the District Court

¶ 58 As previously noted, the district court reasoned that the suppressed transcripts contained non-cumulative impeachment evidence favorable to Tillman because (1) Sgt. Thirsk displayed a great deal more disbelief and incredulity at Sagers's account of the murder in the transcripts than he displayed at trial, (2) some of Sgt. Thirsk's statements and questions in the transcripts gave the appearance that he was coaching Sagers into giving more believable testimony, and (3) the partial transcripts contained over sixty notations indicating that Sagers laughed while being questioned about the murder. The State challenges the district court's conclusion. It argues that the first two categories of non-cumulative evidence identified by the district court, relating to Sgt. Thirsk's disbelief and the appearance of coaching, were cumulative of other evidence available at trial, and therefore should not have been considered in the district court's Brady materiality analysis. The State also challenges the district court's conclusion that the notations of Sagers's laughter in the transcripts reflected inappropriate levity. We examine each of these arguments in turn.

1. Evidence of Sgt. Thirsk's Disbelief and Incredulity

¶ 59 In its ruling, the district court observed that, during trial, Sgt. Thirsk testified that "at one point" during a conversation with Sagers, "I told her that I did not believe her answers to my questions and told her that I believed she had in fact struck [Schoenfeld]." (Emphasis added.) According to the district court, that isolated statement at trial stands in sharp contrast to evidence contained in the suppressed transcripts, in which Sgt. Thirsk's statements and questions indicate that he believed Sagers was "excluding details or not telling a credible story" in "[n]early every page of the first partial transcript ... and many of the pages of the second partial transcript." The court reasoned that defense counsel could have used these dramatic displays of incredulity to undermine the assertion that Sgt. Thirsk had made a "one-time display of incredulity" about Sagers's account, thereby attacking Sagers's credibility and increasing the possibility that the jury may have chosen not to impose the death penalty.

¶ 60 The State argues that the district court erred in concluding that the additional information contained in the transcripts relating to Sgt. Thirsk's displays of incredulity and disbelief provided new information otherwise unknown to the defense. It argues, in essence, that Sgt. Thirsk's displays of incredulity in the transcripts were merely cumulative of Sgt. Thirsk's statement in front of the jury that he "did not believe [Sagers's] answers to [his] questions." Although we agree that this evidence was cumulative of other evidence available to defense counsel prior to trial, we do so for a different reason than that asserted by the State.

¶ 61 Our examination of the record reveals that, contrary to the district court's observation, Sgt. Thirsk did not testify as to simply a one-time display of disbelief during trial. Rather, while Sgt. Thirsk was examined outside the presence of the jury regarding the results of Sagers's polygraph tests, he confirmed, in response to defense counsel's inquiry, that prior to trial he had told other investigators, an individual from the county attorney's office, and both defense counsel that he didn't believe Sagers "for one minute." Because it is evident that defense counsel already knew Sgt. Thirsk greatly distrusted Sagers's account of the murder, and because Sgt. Thirsk explicitly testified as to that fact, the evidence of Sgt. Thirsk's disbelief in the suppressed transcripts is simply cumulative of other evidence available to defense counsel prior to trial. Consequently, the district court erred in concluding that Sgt. Thirsk's displays of incredulity in the transcripts were material and prejudicial for purposes of Brady.

2. Evidence of Coaching by Sgt. Thirsk

¶ 62 In its ruling, the district court acknowledges that Sgt. Thirsk's skepticism in the transcripts about Sagers's account of the murder "was coupled with encouragement to tell the truth." Nevertheless, the court concluded that some of Sgt. Thirsk's statements and questions suggested that he was encouraging Sagers to "simply tell a story that could be believed," and that this evidence of coaching could have been used by the defense to undermine Sagers's credibility. Examples of Sgt. Thirsk's coaching in the suppressed transcripts are as follows:

Q: I understand the defense council [sic] is not going to be very friendly. And he is'nt [sic] going to be saying Gee [sic] Carla how did you know where the light was.

A: Yea I know that.

Q: She's going to hit you with it like that [sic] how did you know where the light was and you know what that applies [sic] to the jury?

A: That I've been there before.

Q: That's right. Or something that adds to [sic] do you see what I mean?

A: ...

Q: How did you know he was going to hit him again? I felt it, nobody buys that.

A: Well I (laugh).

Q: Ok, nobody buys that. Even if it's true nobody buys it, something has to have been happened [sic] either within your visual,

A: Well why else would he have to cover his head up and

Q: Don't ever say that. Why else would he, it does'nt [sic] answer a question.

A: I know it does'nt [sic] but ...

(Ellipses in original.)

¶ 63 On appeal, the State challenges the district court's conclusion that the transcripts evidence "coaching" at all, suggesting instead that the transcripts reflect only "aggressive interrogation." The State also argues that, even if the transcripts give the appearance of coaching, the issue of coaching was "nothing new." In support of this argument, the State observes that, at trial, defense counsel questioned Sgt. Thirsk about possible coaching in the following manner:

Q. Again directing your attention to the January 3rd conversation, did you ever indicate to Miss Sagers that you wanted her to say or testify that she had hit [Schoenfeld]?

A. No, sir, I did not.

Q. Did you ever indicate or imply to Miss Sagers that you wanted her to testify falsely?

A. No, sir, I admonished her about doing that.

As a result, the State argues that any coaching evidence contained in the transcripts would have been merely cumulative of other evidence presented at trial.

¶ 64 The problem with both parties' arguments is that, regardless of whether Sgt. Thirsk was engaged in "aggressive interrogation" or "coaching," the suppressed transcripts arguably present the appearance that Sgt. Thirsk was coaching Sagers by encouraging her to tell a more believable story. Furthermore, contrary to the State's assertions, the transcripts would, in fact, have provided defense counsel with evidence that was not otherwise available either before or during trial—namely, a means by which defense counsel could have attempted to rebut Sgt. Thirsk's denial that he coached Sagers.

¶ 65 The State counters that, even if the transcripts did contain evidence of coaching that was otherwise unavailable to the defense, such evidence would actually have been damaging to Tillman because Sgt. Thirsk's questioning "caused Sagers to admit more involvement in the case, not less." According to the State, further emphasizing the possibility of coaching "may have caused the jury to believe that Sagers was even less involved, and that she only confessed to more involvement because [Sgt.] Thirsk bullied her into it."

¶ 66 While both Tillman and the State put forward plausible theories addressing the potential impact of the coaching evidence, we are persuaded that Tillman would have benefitted if the evidence had been presented at trial. For, even if the transcripts indicated that Sagers could be coached into admitting more involvement about certain aspects of the murder, Tillman could still have used the transcripts to argue to the jury that other aspects of her testimony regarding her own role and Tillman's role in the murder were either fabricated or exaggerated as well.

¶ 67 Additionally, contrary to the State's suggestion, the possible coaching reflected in the transcripts does not involve Sagers simply admitting greater involvement in the murder. As the district court accurately concluded, the transcripts also evince an overall attempt on the part of Sgt. Thirsk to get Sagers to present a more believable or credible narrative of the murder to the jury. Had the transcripts been properly disclosed prior to trial, Tillman could have, at the very least, introduced, in the penalty phase, specific instances of possible coaching contained in the transcripts to explore the extent to which Sagers's testimony changed or was crafted as a result of Sgt. Thirsk's questioning. This may have raised questions in the minds of the jurors as to the overall veracity or credibility of Sagers's account of her and Tillman's respective roles in the murder and perhaps affected a juror's assessment of the appropriate punishment to impose on Tillman.

¶ 68 Thus, we agree with the district court that the evidence of possible coaching in the transcripts was not cumulative, and could have been used to benefit Tillman during the sentencing phase of his trial. Like the other non-cumulative evidence identified above, we leave for later analysis whether, when considered in combination with all the other non-cumulative evidence contained in the suppressed transcripts, this evidence of coaching undermines our confidence in the sentence imposed. We now turn to the State's final evidentiary challenge regarding the suppressed transcripts' numerous notations of Sagers's laughter.

3. Sagers's Laughter

¶ 69 In its ruling, the district court found that the suppressed partial transcripts contain approximately sixty notations indicating moments when Sagers laughed while being questioned by Sgt. Thirsk about the murder. The district court found that while "some of these notations refer to nothing more than nervous laughter," others "appear inappropriate in circumstances where the violent death of a human being and [Sagers's] involvement in that death are being discussed." For example,^[12] while Sagers was explaining that she turned her back while Tillman struck Schoenfeld in the head, the following exchange occurred:

A. I didn't want to see it.

Q. You didn't see ... wanna see what?

A. I didn't want to see him getting hit.

Q. How did you know he was going to be hit?

A. (laugh)

Then, while Sgt. Thirsk was attempting to reconstruct the events surrounding the murder, Sagers laughed at the fact that she could not account for the location of the ax used to bludgeon Schoenfeld:

Q. O.k. where's the hatchet? You just used both hands to cover my head.

A. (laugh)

Similarly, while responding to Sgt. Thirsk's questioning regarding the development of the plan to kill Schoenfeld, the following exchange occurred:

Q: How was he going to do it?

A: Just knock him out when he walked in the door.

Q: Well did he say before that he was going to wait until he was gone then wait for him inside or how did he say he was going to ya know accomplish this feat?

A: (laugh).

¶ 70 The district court concluded Sagers's exhibition of such flippancy, "while in her immunized situation, about the homicide and the role she played in carrying it out certainly would have assisted trial counsel in painting a picture of [her] as someone who was not to be believed."

¶ 71 The State challenges the district court's conclusions regarding Sagers's laughter in two respects. First, it argues that it is just as likely, and perhaps even more likely, that the notations of laughter contained in the transcripts simply indicate a nervous response to being interrogated, rather than an inappropriate reaction to the questions themselves. However, in advancing this argument, the State does not challenge as clearly erroneous the district court's factual finding that the laughter was inappropriate, and we therefore accept that at least some of the levity displayed by Sagers during questioning was inappropriate. See State v. Widdison, 2001 UT 60, ¶ 60, 28 P.3d 1278 (noting that a party who wishes to challenge a factual finding must first marshal the evidence in support of the finding and then show why the marshaled evidence fails to support the finding). Moreover, even if the State had challenged the district court's finding, we would be disinclined to conclude that the district court's finding of inappropriate levity was clearly erroneous.

¶ 72 During the evidentiary hearing held by the district court to consider Tillman's petition for post-conviction relief, Sgt. Thirsk testified that, during his contact with Sagers, she "tended to laugh a lot or giggle a lot," so it was "something [he] noted." When asked how much he recollected about Sagers's laughter, Sgt. Thirsk responded in this way:

A. Not much. I remember at the time that it bothered me.

Q. Why did it bother you?

A. Because it seemed out of character for the seriousness of the situation.

Q. Did it seem as if she was taking the process seriously?

A. At times, yes; at times, no.

Q. Did that laughter seem like it was nervous laughter, or did it seem more lighthearted?

A. Both.

Q. And if we were to focus on using the word "lighthearted," what would that mean to you?

A. Something both of us thought was comical or odd, maybe in the discussion, the pre-test, about things of other subjects.

Q. Was she laughing at things you thought maybe she shouldn't have been laughing at?

A. A couple of times, yes.

The fact that Sgt. Thirsk remembered Sagers's laughter almost twenty years after the interviews in question and that the laughter "seemed out of character for the seriousness of the situation" and included lighthearted laughter at inappropriate times strongly supports the district court's finding that the notations of laugher were more than simply a nervous response.

¶ 73 Even if Sagers exhibited inappropriate levity during her interviews with Sgt. Thirsk, the State argues that the district court's conclusion that such laughter would have undermined Sagers's credibility "is a leap of speculation not supported by any authority or evidence." It argues that, "[i]f counsel were able to convince the jury that [Sagers's] laughter were inappropriate, it might establish that [Sagers] was a more reprehensible person, but it would not establish that she was a liar" and would therefore not affect the credibility of her description of the events that occurred the night of the murder.

¶ 74 Tillman counters that the evidence of Sagers's laughter during her interviews would have affected the jury's assessment of her credibility. He argues that her flippancy, as documented in the suppressed transcripts, could have been interpreted as rebutting the veracity of her claims during trial that she played only a minor role in killing Schoenfeld and was intimidated by her fear of Tillman into cooperating with the murder. Tillman asserts that undermining the State's portrayal of Sagers as a victim who was dominated by Tillman and pressured into participating in the murder would have benefitted him by directly affecting the jury's calculation of Tillman's future dangerousness during the penalty phase of the proceeding. We agree.

¶ 75 The notations of laughter raise questions about the credibility of Sagers's story and therefore constitute valuable impeachment evidence. The inappropriate levity Sagers displayed in the suppressed transcripts could be viewed as evidence of Sagers's inability to grasp the gravity of her situation, leading jurors to question whether Sagers understood the overriding importance of providing completely truthful testimony. Such levity may have also raised questions as to whether Sagers was willing to modify her testimony to insulate herself from blame, protect her immunity from prosecution, and respond to attendant pressures demanding a credible narrative of the events surrounding Schoenfeld's murder.^[13]

¶ 76 Evidence demonstrating Sagers's inappropriate levity would also have proven valuable to Tillman in his efforts to crack the veneer of helplessness and victimization within which the prosecution took pains to encase Sagers at trial. Tillman could have used the frequently flippant attitude displayed by Sagers during the interviews to counteract the prosecution's presentation of Sagers as another one of Tillman's victims, who only participated in the murder plot because she was intimidated by Tillman and fearful of reprisals if she chose not to cooperate.

¶ 77 The notations of laughter contained in the suppressed transcripts are unlike any other evidence that was available to Tillman before or during his trial. As a result, we agree with the district court that evidence of Sagers's laughter is not merely cumulative of other evidence.

¶ 78 Having identified the relevant non-cumulative evidence contained in the suppressed transcripts, we next analyze of whether this evidence, considered collectively, is sufficient to undermine confidence in Tillman's death sentence. For the reasons discussed below, we agree with the district court that confidence in Tillman's death sentence is undermined in light of the suppressed evidence. As a result, we conclude that the suppressed evidence is material for Brady purposes.

C. When Considered Collectively, the Suppressed Evidence Undermines Confidence in Tillman's Death Sentence

¶ 79 As discussed above, the State concedes that the first two prongs^[14] of Tillman's Brady claim are met in this case. Therefore, we confine our inquiry to a determination of whether the absence of the suppressed transcripts prejudiced Tillman during the sentencing phase of his trial. We conclude that Tillman was prejudiced by the absence of the evidence and therefore affirm the district court's decision to vacate his sentence and order a new sentencing hearing.

¶ 80 In order to determine whether the absence of suppressed evidence favorable to the defense resulted in prejudice to a defendant, the evidence in question must be "material." See Strickler, 527 U.S. at 282-83, 119 S. Ct. 1936. A defendant can successfully establish materiality by showing that "there is a reasonable probability that, had the evidence been disclosed to the defense, the result of the proceeding would have been different." Kyles, 514 U.S. at 433, 115 S. Ct. 1555 (internal quotation omitted). A sufficiently reasonable probability of a different outcome is established "when the government's evidentiary suppression undermines confidence in the outcome." Id. at 434, 115 S. Ct. 1555 (internal quotation omitted). When undertaking a materiality analysis, the suppressed evidence must be evaluated in light of the entire record. Agurs, 427 U.S. at 112, 96 S. Ct. 2392. An "item by item" evaluation of the suppressed evidence may be conducted, but it is the cumulative or collective effect of the suppressed evidence that ultimately must be weighed to determine whether confidence in a particular outcome is undermined by the absence of the suppressed evidence. See Kyles, 514 U.S. at 436 & n. 10, 115 S. Ct. 1555.

¶ 81 In accordance with the standard outlined above, we must determine whether the absence of the following evidence at Tillman's trial undermines confidence in the death sentence imposed by the jury: (1) the dramatic improvement, just before trial, in Sagers's ability to recall the sequence of events leading up to and following Schoenfeld's murder, (2) the appearance of coaching by Sgt. Thirsk, (3) the numerous notations of laughter indicating inappropriate levity on the part of Sagers, and (4) Sagers's statements that Tillman was depressed and suicidal prior to the murder.

¶ 82 The State argues that the suppressed evidence is too inconsequential to justify disturbing Tillman's death sentence. Tillman disagrees, arguing that the presence of the suppressed evidence at trial would have significantly undercut the prosecution's efforts to justify the imposition of the death penalty. In essence, Tillman asserts that the ability to more effectively attack Sagers's credibility and increase her moral culpability in the eyes of the jury would have had a positive impact on his sentencing process.^[15] The State, on the other hand, argues that Sagers's credibility and moral culpability are not relevant to a determination of whether Tillman deserves the death penalty.

¶ 83 The district court agreed with Tillman, stating that "the less credible [Sagers's] testimony was shown to be, the greater her degree of moral culpability would have been in the minds of the jurors. Moreover, the greater her moral culpability, the less likely jurors would have voted to impose a sentence of death on [Tillman]." The district court did not explain why it concluded that a death sentence would be less likely for Tillman if Sagers's moral culpability were increased in the minds of the jurors. On appeal, the State argues that there is no rational explanation for the district court's pronouncement. We disagree.

¶ 84 A critical component of the prosecution's argument justifying the imposition of the death penalty on Tillman was that Sagers herself, like Lori Groneman before her, was one of Tillman's victims. The success of this strategy depended heavily on the prosecution's ability to diminish, to the greatest extent possible, any moral culpability on the part of Sagers. The prosecution attempted to accomplish this goal by painting Sagers as a submissive, innocent young woman who was swept away by the deviousness of Tillman. During the guilt phase of Tillman's trial, the prosecution stated that Sagers "got subjected into this situation because of the fact that she was a virgin, because of the fact that she felt guilty about an abortion, and all the rest." At the penalty phase of Tillman's trial, the prosecution's theme was reprised. The image of Tillman corrupting Sagers was once again brought to the jury's attention, with the prosecution stating that Tillman led Sagers "to believe one thing when ... another was occurring. And I think you look at Carla Sagers and the devastation that has occurred to her life and you say to yourself, ``There is a lot of victims [sic] in this case.'"

¶ 85 The prosecution painted Tillman as a man who repeatedly manipulated and took advantage of innocent women, with Sagers merely serving as his latest victim. The prosecution attempted to convince the jury of this by analogizing Sagers to other women in Tillman's life. For example, the State pointed out that Lori Groneman became sexually involved with Tillman when she was an impressionable seventeen. Although Sagers was thirty when her own sexual relationship with Tillman commenced, the prosecution pointed out that it was Sagers's first sexual experience and stated that "[s]exual relationships are very serious business, especially for someone like Carla." The prosecution also drew parallels between Sagers and Tillman's wife. "[C]onsider Carla Sagers, who interestingly enough sort of reminded me a little bit of Mr. Tillman's wife, led along, led to believe one thing when in fact another was occurring." By drawing these parallels, the prosecution was strongly implying that Tillman engaged in a pattern of behavior destructive to the women he encountered—a pattern, the prosecution suggested, that could only be terminated by Tillman's execution.

¶ 86 As the above excerpts show, the efforts to reduce Sagers's moral culpability were not undertaken solely to boost her credibility on the witness stand. Those efforts were also an attempt to justify the imposition of the death penalty on Tillman and to alleviate concerns in the jurors' minds about the potential for drastic disparity in punishment if Tillman received the death penalty while Sagers, pursuant to her immunity grant, walked away completely unpunished. The prosecution directly acknowledged the disparity likely weighing on the jurors' minds, stating that the jurors may have asked themselves, "if we execute or find the sentence of death against Mr. Tillman, shouldn't we also find one against Carla Sagers?" The prosecution chose to address that concern by taking great pains not only to relieve Sagers from moral responsibility for the murder, but to actually cast her as another one of Tillman's victims.

¶ 87 The suppressed evidence that was unavailable to Tillman during his trial could have been used to undermine the prosecution's attempts to diminish Sagers's moral culpability and present her as another in a long line of Tillman's victims. The suppressed transcripts reveal that, shortly before trial, Sagers experienced a marked improvement in her ability to confidently recall the sequence of events surrounding Schoenfeld's murder. Additionally, the transcripts contain numerous passages in which Sgt. Thirsk appears to be coaching Sagers to enable her to supply a more believable narrative. The evidence of Sagers's dramatic improvement of memory and the appearance of coaching on the part of Sgt. Thirsk directly affect Sagers's credibility as a witness, which may have led the jury to believe that Sagers was more involved in the murder than her testimony implied. Tillman could have utilized this evidence to portray Sagers's testimony as a work in progress, carefully honed by the prosecution over the course of many months, and which only took its final shape mere days before trial. The information contained in the suppressed transcripts would have helped Tillman advance the argument that Sagers's testimony was forged in the heat of Sgt. Thirsk's interrogations and was motivated by a desire to please the people who had granted her complete immunity.

¶ 88 Similarly, Tillman could have utilized evidence of Sagers's inappropriate levity during her interactions with Sgt. Thirsk not only to attack her credibility, but to more directly confront the efforts by the prosecution to present her as just one more of Tillman's many victims. The State concedes that the notations of laughter could have led the jury to consider Sagers a more reprehensible person, but then asserts that any moral assessment of Sagers is irrelevant to Tillman's sentencing proceeding. We reject the State's argument that there is no connection between the jurors' assessment of Sagers and their assessment, for sentencing purposes, of Tillman. Evidence tending to undermine Sagers's status as an innocent young woman corrupted by Tillman would have aided Tillman during the penalty phase of his trial. The more morally reprehensible Sagers appeared to jurors, the less inclined they would be to view her as a victim. Further, increased moral culpability on the part of Sagers would throw into even harsher relief the disproportionate treatment of Tillman when compared with the treatment of Sagers. The record reveals that the prosecution itself was concerned that the disparity in treatment would potentially disincline the jury to impose the death penalty, and any evidence tending to widen that disparity would have likely affected the jury's deliberations in the penalty phase.

¶ 89 We express no opinion as to whether, as a general matter, it is appropriate for juries to consider punishments imposed on co-defendants or accomplices when determining whether the death penalty is appropriate. This case involves a unique situation in which the prosecution attempted to present Sagers not as an accomplice to murder, but as a victim of the capital defendant. This characterization made Sagers's moral culpability for the crime highly relevant to the ultimate sentencing determination because, if the defense could show moral culpability on the part of Sagers, it would undermine a critical justification for the imposition of the death penalty. The State's grant of full immunity to Sagers, while obviously an indication of the value of her testimony, can also be viewed as an attempt to bootstrap Sagers into the status of an innocent victim, undeserving of punishment. In this sense, the grant of immunity is just another marker of Sagers's alleged status as "victim," a status designation that Tillman was certainly entitled to attack as part of his defense.

¶ 90 Finally, the defense was unable to take advantage of Sagers's express statement in the suppressed transcripts that Tillman was depressed and even suicidal prior to Schoenfeld's murder. At the time of Tillman's trial, section 76-3-207(2) of the Utah Code listed several non-exclusive mitigating factors to be considered by juries during the sentencing phase of trials. Included in that list was that the "murder was committed while the defendant was under the influence of extreme mental or emotional disturbance." Utah Code Ann. § 76-3-207(2)(b) (Supp. 1982) (current version at Utah Code Ann. § 756-3-207(4)(b) (2003)); see also id. § 76-3-207(2)(g) (allowing capital defendants to present "any other fact in mitigation of the penalty" during the sentencing phase of trial) (current version at Utah Code Ann. § 76-3-207(4)(g) (2003)). If information about Tillman's depression and suicidal thoughts prior to Schoenfeld's murder had been presented to the jury, it may have affected sentencing deliberations.

¶ 91 We are convinced that, when considered collectively and in light of the entire record, the evidence contained in the suppressed transcripts is of sufficient import that its absence at trial undermines confidence in the sentence imposed on Tillman. We note that the sentencing phase of a capital trial is not a scientific process, but rather requires the weighing of a multitude of both aggravating and mitigating factors. See id. § 76-3-207(2) (listing six types of mitigating circumstances and concluding that a defendant may present "any other fact in mitigation of the penalty"); California v. Brown, 479 U.S. 538, 545, 107 S. Ct. 837, 93 L. Ed. 2d 934 (1987) (O'Connor, J., concurring) ("The individualized assessment of the appropriateness of the death penalty is a moral inquiry... [and] should reflect a reasoned moral response to the defendant's background, character, and crime...."); State v. Wood, 648 P.2d 71, 79-80 (Utah 1982) (stating that no specific weight is to be assigned to individual aggravating and mitigating factors and that all circumstances must be weighed when considering the appropriateness of the penalty).

¶ 92 We also note that imposing the death penalty on Tillman required a unanimous decision on the part of the jury. Utah Code Ann. § 76-3-207(3). Consequently, all that was necessary for Tillman to avoid the death penalty was a single doubting juror. If just one juror, while considering the aggravating and mitigating factors, had concluded that the death penalty was inappropriate under the circumstances, a punishment of life imprisonment would have been imposed. If the suppressed transcripts had been available to Tillman, he could have more effectively countered the prosecution's attempts to add Sagers to the list of his victims. That ability may very well have been the difference between life and death.^[16] While the suppressed transcripts do not contain any earthshattering revelations, they do contain significant evidence that damages the credibility of the prosecution's star witness and undermines critical aspects of the prosecution's theory as to why the death penalty was justified in this case. We are not certain that the disclosure of the transcripts prior to Tillman's trial would have resulted in a different sentence. We are, however, compelled to conclude that there exists a significant possibility that Tillman would have achieved a more favorable sentence had the State fully complied with its constitutionally mandated disclosure obligations. That possibility undermines our confidence in the sentence imposed, and we therefore affirm the district court's decision to vacate that sentence and commence a new sentencing hearing.

CONCLUSION

¶ 93 Tillman has demonstrated good cause excusing his failure to raise his current Brady claim in an earlier petition for post-conviction relief. The suppression of the transcripts giving rise to his current claim, coupled with affirmative representations by the State that no such transcripts existed, lead us to conclude that Tillman's current claim was overlooked in good faith and that his current petition was not motivated by an intent to delay or otherwise abuse the post-conviction relief process.

¶ 94 Tillman has also demonstrated that his right to due process was violated by the absence of the suppressed transcripts at his trial. The evidence contained in those transcripts tends to undermine the credibility of Sagers, unquestionably the most important trial witness, and undercuts the prosecution's attempts to paint Sagers as an innocent victim. The suppressed evidence also includes an express statement by Sagers that Tillman was suicidal prior to the time he committed the murder. Considered collectively, the evidence contained in the suppressed transcripts would have aided Tillman's efforts to mitigate his moral culpability while highlighting the moral culpability of Sagers. This, in turn, would have called into question Sagers's status as another victim of Tillman and would also have accentuated the State's dramatically disparate treatment of Tillman and Sagers with respect to punishment, making the imposition of the death penalty less likely. Given the significance of the suppressed evidence and the fact that swaying just one juror would have resulted in a more favorable sentence for Tillman, we conclude that Tillman is entitled to a new sentencing proceeding. We therefore affirm.

¶ 95 Chief Justice DURHAM, Associate Chief Justice WILKINS, Justice PARRISH, and Judge KAY concur in Justice DURRANT'S opinion.

¶ 96 Having disqualified himself, Justice NEHRING does not participate herein; District Judge THOMAS L. KAY sat.

NOTES

[1] Tillman first directly appealed his conviction, which we affirmed in State v. Tillman, 750 P.2d 546 (Utah 1987). Tillman then filed a state petition for habeas corpus relief, which the district court denied. We affirmed that denial in Cook, 855 P.2d 211. After a petition for federal habeas relief was dismissed for failure to fully exhaust state remedies, Tillman filed a second petition for state habeas relief directly with this court, which we dismissed as procedurally barred. Tillman then filed a second petition for federal habeas relief, which the federal district court denied. Tillman v. Cook, 25 F. Supp. 2d 1245 (D.Utah 1998). The United States Court of Appeals for the Tenth Circuit affirmed the district court's denial, Tillman v. Cook, 215 F.3d 1116 (10th Cir.2000), and the United States Supreme Court denied Tillman's petition for certiorari on December 11, 2000, Tillman v. Cook, 531 U.S. 1055, 121 S. Ct. 664, 148 L. Ed. 2d 566 (2000).

[2] In its brief, the State suggests that one of the interviews evidenced by the partial transcripts could have been the first polygraph interview Lieutenant Bill Robinson conducted with Sagers shortly after Schoenfeld's murder in May 1982. However, the district court found that the interviews in question were those conducted by Sgt. Thirsk, and the State has failed to challenge that finding as clearly erroneous. Therefore, we accept the district court's finding that the two transcripts contain portions of the interviews of Sagers conducted by Sgt. Thirsk. See State v. Widdison, 2001 UT 60, ¶ 60, 28 P.3d 1278 (noting that a party who wishes to challenge a factual finding must first marshal the evidence in support of the finding and then show why the marshaled evidence fails to support the finding).

[3] The tapes from which the interviews with Sagers were transcribed have never been provided to the defense.

[4] Tillman also argued that he was entitled to a reduced sentence of life imprisonment because the prosecutor improperly injected religion into his closing argument during the penalty phase of trial. The district court dismissed this argument as procedurally barred, and it is not before us on appeal.

[5] As an initial matter, we observe that Tillman has included various extra-record material in his appellate brief and accompanying addendum. Because our review is limited to only those materials contained in the record, see Wilderness Bldg. Sys., Inc. v. Chapman, 699 P.2d 766, 768 (Utah 1985), we disregard any improperly included materials.

[6] Utah Code Ann. §§ 78-35a-101 to -304 (2002 & Supp.2004).

[7] To avoid confusing the "reasonable probability" standard with the more demanding "more likely than not" standard, Justice Souter has advised courts to view "reasonable probability" as more akin to a "significant possibility" of a different result. See Strickler, 527 U.S. at 297-300, 119 S. Ct. 1936 (Souter, J., concurring and dissenting).

[8] Prior to trial, Tillman was apparently provided with three transcripts in which Sagers made numerous statements to prosecutors and police, as well as a transcript of the testimony Sagers gave at the preliminary hearing. Only the preliminary hearing transcript is contained in the record on appeal. Consequently, while Tillman has attached the three remaining transcripts in the addendum to his appellate brief, we do not rely on them for purposes of our analysis.

[9] Because the transcripts are not dated, it is unclear whether the partial transcript upon which Tillman relies was from Sagers's interview in December 1982, approximately four weeks before Tillman's trial began, or from her interview in January 1983, just days before trial.

[10] In her preliminary hearing testimony, Sagers stated that she did not turn Tillman in to authorities because he was a "repentful" person. Tillman acknowledges that the word "repentful" may actually have been "revengeful" and that the court reporter may have mistyped the word. If such a transcription error occurred, Tillman's argument that Sagers characterized Tillman as "revengeful" for the first time at trial would clearly lack merit. However, even if the transcript is accurate, it only further illustrates that Tillman had evidence in his possession prior to trial that could have been used to impeach Sagers's trial testimony that she refrained from contacting police due to her fear of Tillman's "revengefulness." As the suppressed transcripts contain no additional evidence that would have meaningfully aided such an impeachment attempt, Tillman cannot successfully claim that he was prejudiced by the absence of the suppressed transcripts in this regard.

[11] As an additional observation, we note that even if we were to rely on the three transcripts not contained in the record, a portion of at least one transcript quoted by Tillman demonstrates that Sagers resisted the assertion that she feared Tillman. Specifically, in an interview with prosecutor Mike Christensen, Sagers expressly denied that Tillman had ever hit her or threatened her or her family with injury and stated that Tillman only got "rough" with her once when, the morning after she refused to shoot Schoenfeld and Groneman, Tillman pushed her out of the doorway to her apartment while retrieving his gun. It would appear that this categorical denial of Tillman's dangerousness could have been used by the defense to undermine Sagers's "revengefulness" testimony at trial far more effectively than the mere absence of statements about Tillman's "revengefulness" in the undisclosed partial transcripts.

[12] Although we quote only three examples of inappropriate laughter, we observe that the transcripts are replete with instances where Sagers appears to exhibit inappropriate levity during questioning by Sgt. Thirsk.

[13] The following excerpt is one of the many examples contained in the suppressed transcripts where a notation of laughter can reasonably be viewed as undermining confidence in the veracity of an answer provided by Sagers:

Q: And when was this about?

A: ... I don't remember when this happened.

Q: You don't have to give me an exact date your best recollection? [sic]

A: ... March.

Q: Ok.

A: (laugh).

[14] The first two prongs of a Brady claim require a showing that evidence favorable to the defense was suppressed by the State. See Strickler, 527 U.S. at 281-82, 119 S. Ct. 1936.

[15] As the district court noted, Sagers's credibility and moral culpability are related to each other. At trial, Tillman attempted to magnify the role Sagers played in Schoenfeld's murder. In furtherance of this goal, Tillman called Sgt. Thirsk, who testified that he believed Sagers participated in the murder to a greater extent than her testimony would suggest. It is highly likely that doubts about Sagers's credibility would lead jurors to assume, just as Sgt. Thirsk did, that Sagers played a greater role in Schoenfeld's murder than her trial testimony suggested.

[16] As noted by Justice Stewart in his dissent in Tillman v. Cook, 855 P.2d 211, 231 (Utah 1993), the imposition of the death penalty on Tillman was not a foregone conclusion:

There were mitigating factors in this case. Tillman's family testified on his behalf. Furthermore, virtually all of the State's case against Tillman came from a person who, herself, could have been prosecuted for capital homicide, and yet was given total immunity from prosecution.... Tillman had no record of any prior homicide, a factor that has been significant in a number of death penalty cases in this state, especially where there was only one homicide. Finally, although Tillman has a criminal record, his most serious crimes were committed more than twenty years ago, and only one, attempted robbery committed in 1962, appeared to be a crime of violence.