BACKGROUND

A. Underlying Facts and Trial

In July 2005, Weeden and three other fourteen-year-old girls were walking down the street when a car approached.1 Inside were four boys, including twenty-two-year-old Deovinesh Kumar and seventeen-year-old Navnil Chand. Kumar invited the girls to “party,” promising beer and “weed.” The girls declined, but one of them, Angela, gave the boys Weeden’s cell phone number. Later that night, Weeden and Angela met with two of their friends, fourteen- year-old John W. and sixteen-year-old Ryan Moore. Weeden told them about the interaction with Kumar and Chand. Moore suggested robbing them; Weeden replied “yeah, we should.”

Days later, Weeden told Angela that the robbery would take place at a park. Weeden asked John to participate, but he refused. Moore told John’s cousin, Janee Hill, of a plan to rob some “East Indian boys” for “weed and money.”

On August 5, Weeden told Hill that the boys she met had been “crank calling” her and inviting her to a motel. Weeden told Hill that Moore was going to rob the boys. Hill warned that robberies can go wrong; Weeden responded “okay.”

That evening, after calling Weeden, Chand told Kumar that the girls they had encountered in July would meet them

1 We rely on the California Court of Appeal’s recitation of the evidence. See People v. Melonson, Nos. C061352, C061800, 2013 WL 1987240, at *1–11 (Cal. Ct. App. May 15, 2013), as modified on denial of reh’g (June 14, 2013). at a park. Kumar drove to the park. While waiting in the car, Chand spoke on the phone to a girl who said, “I’ll be there in two to three minutes.” Chand replied, “I’m waiting for you over here by the park.”

At the same time, Moore was with twenty-year-old Sirtice Melonson at a nearby park. Moore was on the phone with Hill, who in turn used a second phone to talk to Weeden. Weeden told Hill to ask Moore whether he saw the boys’ car; Moore said no. Weeden directed Moore to a different park, where Moore saw a gold car. Weeden said “that’s them.”

Melonson approached the car, stuck a handgun in the window, and ordered Kumar and Chand to exit. Before they could comply, the gun went off. Kumar sped off amid further gunfire. Chand later died of a gunshot wound.

Phone records from that evening confirm that: (1) Chand called Weeden repeatedly until the shooting, (2) Weeden and Moore exchanged calls until approximately twenty minutes before the shooting, (3) Weeden was on the phone with Hill until the shooting, and (4) Hill called Moore multiple times before and after the shooting. Neither Hill nor Weeden was near the park. Months later, Weeden sent text messages to Moore acknowledging that she knew about the robbery plan in advance but denying that it was her idea.

The only defense evidence at trial came from four character witnesses; each testified that Weeden was not the sort of person who would plan a robbery. In closing argument, Weeden’s counsel asserted that Melonson decided to commit the robbery without Weeden’s knowledge. He also urged the jury to consider Weeden’s age and manipulability:

WEEDEN V. JOHNSON 7

Try to put glasses on of a 14-year-old girl. . . . You’re easily manipulated. . . . But the Prosecution wants to charge her as an adult, under adult standards. But you can take into consideration, put on those glasses of a 14- year-old girl and tee, hee, hee, older boys want to talk to me and what should I do and let’s send ‘em on a wild good [sic] chase. That’s a 14-year-old girl, it just is, and you get to put those glasses on. And it’s [sic] easily manipulated by older people when she stepped into this mess of vipers. And blaming her for this is like blaming a child molest [sic] victim. She’s 14. She don’t [sic] know what she’s doing. She gets manipulated into sending these boys out to get stood up.

The jury found Weeden guilty of attempted robbery and first- degree felony murder, but acquitted her of attempted murder. Weeden was sentenced to twenty-nine years to life in prison.

B. The New Trial Motion

Represented by new counsel, Weeden moved for a new trial, claiming that her trial counsel was ineffective for failing to investigate or present psychological evidence. In support of the motion, Weeden submitted a psychological report from Lisa Perrine, Ph.D., who evaluated Weeden after the verdict. Dr. Perrine opined that although Weeden could comprehend the concept of robbery, “it is extremely unlikely she would intend to commit robbery or knowingly participate in one,” and “she would probably be slow to understand that a robbery was being considered by others if their intentions were not clearly articulated.” Dr. Perrine also found Weeden to be “quite passive and vulnerable to being manipulated by others,” and concluded she had “serious cognitive deficits (for example, 91% of people her age function[ed] intellectually at a higher level),” “well below average language skills,” and “a strong tendency to miss important environmental cues.”

Weeden’s trial counsel stated in a declaration that he did not investigate psychological testimony because he “did not consider a psychologist’s opinion to be relevant to the issues in this prosecution.” At an evidentiary hearing, trial counsel admitted that he had “contemplated” seeking a psychological evaluation, but did not do so because “regardless of what the doctor would have concluded, it would be inconsistent with the defense that I was putting forth.” Counsel also speculated the prosecution might have used the results of an examination against Weeden:

Q: You didn’t know what the doctor would conclude, though. Is that right?

A: No, that is true. That is true. But – for instance, if the doctor were to conclude that she was completely immature for instance and – and easy – easily manipulated, my concern would have been that the Prosecution would have used that to show that she didn’t understand perhaps the magnitude of a robbery, but still participated in it. So I had some concerns that that could be twisted and turned against her.

WEEDEN V. JOHNSON 9

The state trial court denied the new trial motion, characterizing counsel’s failure to obtain a psychological examination as a “sound tactical decision.”

C. Direct Appeal

The California Court of Appeal affirmed Weeden’s conviction. The court disposed of Weeden’s ineffective assistance claim in four paragraphs. The first paragraph recited the familiar legal standard for establishing ineffective assistance of counsel set forth in Strickland v. Washington, 466 U.S. 668, 687–688 (1984). People v. Melonson, Nos. C061352, C061800, 2013 WL 1987240, at *21 (Cal. Ct. App. May 15, 2013), as modified on denial of reh’g (June 14, 2013). The second paragraph concluded that counsel made a “reasonable tactical decision” not to seek a psychological examination because an examination “might undermine” his trial strategy:

Weeden contends trial counsel’s decision not to have her examined by a psychologist was not a sound tactical decision. However, trial counsel articulated a very reasonable tactical decision for not requesting such an examination. If the examination revealed Weeden was easily manipulated, the prosecution could claim this was evidence that although Weeden did not understand the magnitude of the robbery, she nonetheless went along with it. In essence, the examination might undermine the defense strategy trial counsel was pursuing at trial. Id. The third paragraph acknowledged Weeden’s argument that Dr. Perrine’s testimony might have made a difference at trial, but, rather than determining whether the failure to present that testimony was prejudicial, the fourth paragraph reiterated the risk that such testimony “might well have undermined” counsel’s trial strategy:

In addition, Weeden argues the lack of psychological testimony on her lack of specific intent ‘undermines confidence in the outcome of this trial.’ According to Weeden, the testimony of Dr. Perrine would have raised a reasonable doubt as to whether Weeden harbored a specific intent to rob. Presented with Dr. Perrine’s testimony, the jury would not have concluded Weeden joined in Melonson’s criminal intent to commit armed robbery.

Again, trial counsel pursued the defense that Weeden did not set up the robbery or agree to it. The testimony Weeden claims counsel was ineffective in failing to present might well have undermined this defense. Among Dr. Perrine’s findings was that Weeden had a strong tendency to be a passive follower, which might have led the jury to conclude she did not initiate the robbery but, sheep-like, went along with it. We find no ineffective assistance.

Id.

WEEDEN V. JOHNSON 11

The California Supreme Court summarily denied Weeden’s petition for review.

D. Federal Habeas Petition

Weeden then filed a 28 U.S.C. § 2254 habeas corpus petition in the United States District Court for the Eastern District of California. That court denied the petition and declined to issue a certificate of appealability. Weeden v. Johnson, No. 2:13-CV-02667-JKS, 2014 WL 6808778, at *16–18, *27 (E.D. Cal. Nov. 26, 2014). A panel of this court granted a certificate of appealability. We have jurisdiction of Weeden’s appeal under 28 U.S.C. § 1291 and § 2253(a).

DISCUSSION

We review the district court’s denial of Weeden’s habeas corpus petition de novo. Lambert v. Blodgett, 393 F.3d 943, 964 (9th Cir. 2004). Under the Antiterrorism and Effective Death Penalty Act of 1996 (“AEDPA”), habeas relief may not be granted “with respect to any claim that was adjudicated on the merits in State court proceedings” unless the state decision was (1) “contrary to, or involved an unreasonable application of, clearly established Federal law, as determined by the Supreme Court of the United States,” or (2) “based on an unreasonable determination of the facts in light of the evidence presented in the State court proceeding.” 28 U.S.C. § 2254(d). We apply AEDPA’s standards to “the state court’s ‘last reasoned decision’ on the claim,” in this case the decision of the California Court of Appeal. Edwards v. Lamarque, 475 F.3d 1121, 1126 (9th Cir. 2007) (en banc) (quoting Ylst v. Nunnemaker, 501 U.S. 797, 803–04 (1991)). A. Deficient Performance

Under the two-pronged framework of Strickland, the Sixth Amendment guarantee of counsel in a criminal proceeding is violated if “counsel’s performance was deficient” and “the deficient performance prejudiced the defense.” 466 U.S. at 687. Our review of the Court of Appeal’s holding that Weeden’s counsel was not deficient is “doubly” deferential, because Strickland requires state courts to give deference to choices made by counsel and AEDPA in turn requires us to defer to the determinations of state courts. Harrington v. Richter, 562 U.S. 86, 105 (2011) (quoting Knowles v. Mirzayance, 556 U.S. 111, 123 (2009)).

Even under this demanding standard of review, we conclude that the Court of Appeal’s conclusion that Weeden’s trial counsel provided effective representation was “contrary to, or involved an unreasonable application of,” clearly established Supreme Court law. 28 U.S.C. § 2254(d)(1). The Supreme Court has repeatedly made plain that counsel has the “duty to make reasonable investigations or to make a reasonable decision that makes particular investigations unnecessary.” Strickland, 466 U.S. at 691; Kimmelman v. Morrison, 477 U.S. 365, 384 (1986); Wiggins v. Smith, 539 U.S. 510, 521 (2003); Hinton v. Alabama, 134 S. Ct. 1081, 1088 (2014).

The Court of Appeal did not identify any “reasonable decision” made by Weeden’s trial counsel that rendered an investigation of psychological evidence “unnecessary.” Strickland, 466 U.S. at 691. Instead, it concluded that the choice not to investigate psychological evidence was a sound “tactical decision” because counsel feared that the results of an expert evaluation might undermine his trial strategy.

WEEDEN V. JOHNSON 13

Melonson, 2013 WL 1987240, at *21. But that analysis “puts the cart before the horse.” Bemore v. Chappell, 788 F.3d 1151, 1167 (9th Cir. 2015), cert. denied sub nom. Davis v. Bemore, 136 S. Ct. 1173 (2016). Counsel cannot justify a failure to investigate simply by invoking strategy. The Supreme Court has squarely rejected the “attempt to justify [a] limited investigation as reflecting a tactical judgment.” Wiggins, 539 U.S. at 521; see also Williams v. Taylor, 529 U.S. 362, 396 (2000) (failure to uncover mitigating evidence was “not justified by a tactical decision” where counsel “did not fulfill their obligation to conduct a thorough investigation of the defendant’s background”). Under Strickland, counsel’s investigation must determine trial strategy, not the other way around. E.g., Bemore, 788 F.3d at 1166–67; Reynoso v. Giurbino, 462 F.3d 1099, 1112 (9th Cir. 2006); Phillips v. Woodford, 267 F.3d 966, 978 (9th Cir. 2001). Weeden’s counsel could not have reasonably concluded that obtaining a psychological examination would conflict with his trial strategy without first knowing what such an examination would reveal. Equally unpersuasive was counsel’s conclusion that the prosecution could have used the results of an examination against Weeden. While a defendant must disclose expert reports she intends to rely on at trial, Cal. Penal Code § 1054.3(a), simply procuring a report does not mean it must be produced, Sandeffer v. Super. Ct., 22 Cal. Rptr. 2d 261, 264 (Ct. App. 1993); Hines v. Super. Ct., 25 Cal. Rptr. 2d 712, 715 (Ct. App. 1993).

The correct inquiry is not whether psychological evidence would have supported a preconceived trial strategy, but whether Weeden’s counsel had a duty to investigate such evidence in order to form a trial strategy, considering “all the circumstances.” Strickland, 466 U.S. at 691. The answer is yes. The prosecution’s felony murder theory required proof that Weeden had “specific intent to commit the underlying felony,” People v. Jones, 98 Cal. Rptr 2d. 724, 727 (Ct. App. 2000), so Weeden’s “mental condition” was an essential factor in deciding whether she “actually had the required mental states for the crime,” People v. Steele, 47 P.3d 225, 240 (Cal. 2002). The Supreme Court has repeatedly noted that the mind of a fourteen-year-old is markedly less developed than that of an adult, see, e.g., J.D.B. v. North Carolina, 564 U.S. 261, 272–73 (2011); Graham v. Florida, 560 U.S. 48, 68 (2010); Roper v. Simmons, 543 U.S. 551, 569–70 (2005), and trial counsel described Weeden as “unusually immature.”2 Given the exculpatory potential of psychological evidence, counsel’s failure to investigate “ignored pertinent avenues for investigation of which he should have been aware.” Porter v. McCollum, 558 U.S. 30, 40 (2009).

Counsel’s performance was deficient because he failed to investigate, a failure highlighted by his later unreasonable justification for it. We do not suggest that counsel must investigate psychological evidence in every case, or even the ordinary case. But the Supreme Court has made clear that some “[c]riminal cases will arise where the only reasonable and available defense strategy requires consultation with experts.” Harrington, 562 U.S. at 106. For the reasons noted above, this was such a case. The Court of Appeal’s finding that counsel rendered adequate performance because he made a tactical decision not to investigate was therefore “contrary

2 Moreover, several witnesses heard Weeden talking about the robbery in advance, rendering counsel’s argument that Weeden was ignorant of the robbery plan quite weak. See Bemore, 788 F.3d at 1167 (counsel’s duty to investigate mental health defense was “especially critical” where “the weaknesses of an alibi should have been known”).

WEEDEN V. JOHNSON 15

to, or involved an unreasonable application of,” clearly established Supreme Court law. 28 U.S.C. § 2254(d)(1).

B. Prejudice

Counsel’s deficient performance justifies relief only if there is a “reasonable probability” that the jury would have reached a different result if adequate representation had been afforded. Strickland, 466 U.S. at 694. A reasonable probability is one “sufficient to undermine confidence in the outcome.” Id. Although counsel’s deficient performance must have had more than a “conceivable effect” on the outcome, it need not have “more likely than not altered” the outcome. Id. at 693. And, because the jury was required to reach a unanimous verdict on each count, the outcome could have differed if only “one juror would have struck a different balance.” Wiggins, 539 U.S. at 537.3

Had the California Court of Appeal found that trial counsel’s performance, even if deficient, did not meet the Strickland “reasonable probability” standard, we would of course be required to afford that determination AEDPA deference. But the Court of Appeal made no such determination. Although the state court acknowledged Weeden’s argument about prejudice, it pretermitted that point, simply reiterating its previous conclusion that counsel’s performance was not deficient:

3 “The Strickland prejudice analysis is complete in itself; there is no place for an additional harmless-error review.” Avila v. Galaza, 297 F.3d 911, 918 n.7 (9th Cir. 2002) (alteration omitted) (quoting Jackson v. Calderon, 211 F.3d 1148, 1154 n.2 (9th Cir. 2000)). Weeden argues the lack of psychological testimony on her lack of specific intent ‘undermines confidence in the outcome of this trial.’ . . .

Again, trial counsel pursued the defense that Weeden did not set up the robbery or agree to it. The testimony Weeden claims counsel was ineffective in failing to present might well have undermined this defense. Among Dr. Perrine’s findings was that Weeden had a strong tendency to be a passive follower, which might have led the jury to conclude she did not initiate the robbery but, sheep-like, went along with it. We find no ineffective assistance.

Melonson, 2013 WL 1987240, at *21. Because the Court of Appeal did not reach the issue of prejudice, we address the issue de novo. Rompilla v. Beard, 545 U.S. 374, 390 (2005) (examining prejudice de novo where state court did not reach the issue); Wiggins, 539 U.S. at 534 (same); see also Porter, 558 U.S. at 39 (examining deficient performance de novo where state court did not reach the issue); Miles v. Ryan, 713 F.3d 477, 489–90 (9th Cir. 2013) (same).

Even assuming, as the dissent suggests, that the Court of Appeal reached the prejudice issue by speculating that Dr. Perrine’s testimony “might have” led the jury to the same result, Dissent at 35, the state court then plainly applied the wrong legal standard. Strickland asks whether there is a “reasonable probability” of a different result, not whether the jury “might have” reached the same result. 466 U.S. at 694. And, when the state court applies an incorrect standard, our

WEEDEN V. JOHNSON 17

prejudice analysis is “unconstrained” by AEDPA deference. Williams v. Taylor, 529 U.S. 362, 405–06 (2000).

After evaluating Weeden, Dr. Perrine concluded that “it is extremely unlikely she would intend to commit robbery or knowingly participate in one,” that “she would probably be slow to understand that a robbery was being considered by others if their intentions were not clearly articulated,” and that she was “quite passive and vulnerable to being manipulated by others.” This testimony from a qualified expert “would have added an entirely new dimension to the jury’s assessment” of the critical issue of Weeden’s mens rea. Vega v. Ryan, 757 F.3d 960, 974 (9th Cir. 2014) (quoting United States v. Kohring, 637 F.3d 895, 905–06 (9th Cir. 2010)). Weeden was not at the crime scene and did not personally commit either the robbery or the murder. The state’s evidence of her intent was therefore presented through testimony by others about what she said and text messages sent months after the events at issue. That evidence did not offer the kind of critical insight into the effect of Weeden’s youth on her mental state that Dr. Perrine’s conclusions provided.4 Moreover, although the jury found Weeden guilty of attempted robbery and first-degree felony murder, she was acquitted of attempted murder. This fragmented verdict is an “additional indicator of prejudice.” Vega, 757 F.3d at 974.

The state surmises that, if Dr. Perrine had testified, it would have presented rebuttal evidence. But the record contains no such evidence. In determining how omitted evidence would have altered the trial, “we may not invent

4 Although Weeden’s counsel urged the jury to consider her age and manipulability in his closing argument, the jury had scant evidentiary basis on which to do so in the absence of expert testimony. arguments the prosecution could have made.” Hardy v. Chappell, 832 F.3d 1128, 1141 (9th Cir. 2016). The state also speculates that the jury would have assigned little weight to Dr. Perrine’s testimony, because her conclusions were based on interviews conducted years after the offense. While that criticism may support a line of cross-examination at retrial, Dr. Perrine’s testimony is uncontradicted on the record before us, and is directly probative of a critical element in the case, Weeden’s mens rea.

Considering all the circumstances of this case, we conclude that, had Dr. Perrine’s opinion been presented to the jury, the probability of a different result is “sufficient to undermine confidence in the outcome.” Strickland, 466 U.S. at 694. Counsel’s deficient performance therefore requires issuance of the writ.

C. The Dissent

We add a brief rejoinder to the concerns of our dissenting colleague. Contrary to her repeated assertions, we do not today adopt a “per se rule.” Dissent at 20, 34–35. We hold only that: (a) in a murder case involving a defendant who was fourteen years old at the time of the crime; (b) in which mens rea was critical given the defendant’s absence from the crime scene and the prosecution’s felony murder theory; (c) where the defendant’s trial counsel declined to investigate psychological testimony because he feared what he might learn;5 (d) where the only expert psychological report in the

5 The dissent points out that Weeden’s trial counsel had “experience” with murder trials and “arguably executed his strategy at trial effectively.” Dissent at 33. At issue, though, is not his previous experience or trial

WEEDEN V. JOHNSON 19

post-conviction record concluded that the defendant was “extremely unlikely” to have formed the requisite mens rea;6 and (e) where the state court nonetheless concluded that the defendant received effective assistance because a psychological examination “might” have undermined counsel’s preselected strategy,7 habeas relief should issue. A contrary conclusion would reduce the Sixth Amendment guarantee of effective representation of counsel to mere “sound and fury, signifying nothing.” William Shakespeare, Macbeth, act 5, sc. 5.

CONCLUSION

We reverse the judgment of the district court and remand with instructions to grant the writ of habeas corpus unless Weeden is timely retried.8

performance in general, but rather his “particular decision not to investigate” psychological testimony. Strickland, 466 U.S. at 691. 6 The dissent’s speculative suggestion that Dr. Perrine’s testimony “likely would have invited the Government to call its own forensic psychologist,” Dissent at 38, is not grounded in the record, and it is well settled that “we do not ‘evaluate all the hypothetical reasons that could have supported’” a state court’s reasoned decision. Cuero v. Cate, 827 F.3d 879, 884 (9th Cir. 2016) (quoting Cannedy v. Adams, 706 F.3d 1148, 1157 (9th Cir. 2013)). 7 The dissent offers additional possible justifications for the state court’s conclusion, including an argument that defense counsel “could have reasonably believed that his time and available resources were better spent” elsewhere. Dissent at 34. But, we review only the state court’s reasoned decision. Cuero, 827 F.3d at 884. 8 Because we conclude that Weeden’s petition should be granted on the ineffective assistance of counsel claim, we do not reach her claim of instructional error. CALLAHAN, Circuit Judge, dissenting:

There is no denying that this case is tragic: a young man’s life was stolen from him, a family was deprived of a son, and four young people made decisions that will dramatically alter the course of their lives. However, this tragedy is no excuse for making bad law, which is exactly what the majority opinion does. In its quest to free a 14-year-old girl from the consequences of a felony murder conviction, the majority disregards the substantial deference federal courts owe to both trial counsel and state courts, and establishes, in essence, a per se rule regarding experts that will call into question many constitutionally-sound state convictions. Because such action is inconsistent with clearly established Supreme Court precedent and the Antiterrorism and Effective Death Penalty Act of 1996 (“AEDPA”), I respectfully dissent.

I. Background

The shooting and law enforcement’s investigation

On August 5, 2005, 22-year-old Deovinesh Kumar and 17-year-old Navnil Chand were shot by 20-year-old Sirtice Melonson at a local park, apparently during a botched robbery. Chand later died from the wounds he sustained. At the time of the shooting, Melonson was joined by 16-year-old Ryan Moore. Kumar and Chand apparently did not know either Melonson or Moore. However, they did know 14-year- old Weeden, whom they had met the previous week when they ran into Weeden, Angela G., and two other teenage girls. On the night in question, Kumar and Chand phoned Weeden twenty-nine times—the last call being placed just minutes before the men were shot. According to Kumar, he and

WEEDEN V. JOHNSON 21

Chand went to the park to meet one of the teenage girls they had met the prior week.

While Weeden was not near the scene, she had been in communication with Moore throughout the night, including up to twenty minutes before the shooting. Additionally, shortly after Weeden received her first call from Chand, Weeden began exchanging calls with 17-year-old Janee Hill, which continued up to and after the shooting. Besides talking to Weeden, Hill also spoke telephonically with Moore off and on during the night in question. None of the phone calls between the parties were recorded.

Following the night in question, Moore, at the behest of law enforcement, sent Weeden a number of text messages accusing Weeden of being responsible for the shooting. In response, Weeden stated: “[D]is ain’t my fucking fault. Sirtice should [sic] have done what he did. He brought you into that by doing that. This shit wasn’t my fault, so don’t even fucking say it was.” Weeden went on to say: “[I]t’s Angie’s fucking fault, dummy. She da one that went to the car and gave them my number.” When Moore claimed that it was Weeden’s idea to rob the “hindus,” Weeden retorted with,“[i]t wasn’t my fucking idea. It was Angie’s, so shut up.” While Weeden admitted to Moore that she knew “Sirtice was going to rob [the men in the car]” and “that someone [had] told [her] to tell [the men in the car] to meet [her] at the park,” she maintained that “[she] didn’t set nuttin’ up.” Weeden also advised Moore that “[i]f [he] want[ed] to get out of this, [he would] have to give up Angie’s name.”

In addition to the pretextual text messages, law enforcement interviewed 14-year-old Angela G., 14-year-old John W., and Hill, all three of whom are related to each other and to Moore. Angela G. apparently stated that, leading up to the incident, Weeden and Moore had talked about robbing the victims. John W. said that, prior to the shooting, he overheard Weeden telling Angela G. that Moore was going to rob someone.

As for Hill, she initially fooled the homicide detectives into believing that she was someone other than who she was, so the night in question was not discussed during her first encounter with law enforcement. Hill lied about her identity because she had a traffic warrant and was afraid that she would be arrested and taken to juvenile hall if she told the detectives who she was. Eventually, Hill did speak with law enforcement a number of times about the incident. At first, Hill indicated that Moore had nothing to do with the shooting and that the only thing Weeden had said to her on the night in question was to call Moore’s phone and ask “if he was there.” Later, however, Hill told law enforcement that Moore had said to her a few days before the shooting that he was planning to rob someone. Hill also stated that, on the night of the shooting, Weeden kept asking Hill to call Moore and pass along certain messages, such as where to go, how many guys he should expect to see, and what kind of car to look for. Despite this, Hill denied knowing that Melonson, Moore, and Weeden were intending to rob the victims because Moore had told her during one of her calls with him that no robbery was going to occur.

Weeden’s trial

In January 2006, Melonson, Moore, Weeden, and Hill were charged with first degree murder for the death of Chand and attempted murder for the shooting of Kumar. Because Melonson had previously been convicted of felony home

WEEDEN V. JOHNSON 23

invasion, he was also charged with being a felon in possession of a firearm. Later, Moore and Hill pleaded guilty to lesser charges. In exchange for Hills’ plea and testimony against the co-defendants, the first degree murder and attempted murder charges against her were dropped and she was allowed to plead guilty to an offense that had a confinement cap of four years. Prior to trial, the People filed an amended information, adding the charge of attempted second degree robbery against Melonson and Weeden.

On October 27, 2008, the joint trial for Melonson and Weeden began. Among other witnesses, the People called John W. John W. initially denied hearing Weeden talk about a robbery. However, in an about-face, John W. later testified on direct examination that he had heard Angie and Weeden planning a robbery. John W. went on to say that the girls had asked him to help with the robbery, but he said no. While on the stand, Angie G. denied ever hearing anyone, including Weeden, talk about a robbery.

The People also introduced call logs from the night in question. To discuss the content of the phone conversations, the People called Hill. On direct examination, Hill deflected the blame from herself and Moore and indicated that Weeden, the youngest of the four defendants and a person who had no criminal background, was the impetus behind the botched robbery. According to Hill, Weeden initially told her that someone kept crank calling her and that she was going to have the person beat up. However, Weeden later allegedly stated that she was going to have Moore rob the boys who were crank calling her and that she and Moore were going to get “weed and money.” Hill claimed that she warned Weeden that “she shouldn’t do the robbery because bad things happen with robberies.” Hill also discussed how Weeden had her relay a number of messages to Moore that ultimately led Moore and Melonson to the victims’ location.

On cross examination, Weeden’s trial attorney impeached Hill on a number of grounds: (1) her plea agreement that reduced her possible sentence from life to 4 years; (2) the multiple lies Hill had told law enforcement when being interviewed about the shooting to protect Moore and herself; and (3) the fact that she had initially lied to law enforcement about who she was because she was afraid that she was going to be taken to juvenile hall for a traffic warrant.

During Weeden’s case in chief, Weeden’s trial attorney called four character witnesses who stated that Weeden was not the type of person to participate in a robbery. Following the presentation of the evidence, the jury was instructed that the first degree murder charge was based on a felony murder theory and that the jury could convict Weeden of murder only if it believed that she intended for the victims to be robbed. During closing arguments, one of Weeden’s trial attorney’s central themes was that the People had failed to show beyond a reasonable doubt that Weeden directed Moore to the victims’ car to rob them, as opposed to some other reason, such as to scare them or have them beat up. Seven days after the close of the case, the jury found Weeden guilty of first degree felony murder and attempted second degree robbery, but not guilty of attempted murder.1 Weeden’s sentencing hearing was scheduled for approximately two months later.

1 In contrast, Melonson was convicted of all charges brought against him two days after the joint trial ended.

WEEDEN V. JOHNSON 25

Post-trial

Following the trial, Weeden released her trial attorney, who had represented her since at least January 2006, and retained new counsel. At the sentencing hearing, Weeden’s new counsel indicated that he intended to file a number of motions, one of which was a motion for a new trial based, in part, on the allegation that Weeden’s trial attorney was ineffective for failing to have Weeden examined by a forensic psychologist. The Superior Court expressed skepticism of Weeden’s new counsel’s claim of ineffective assistance, stating that the trial attorney’s performance was “excellent” and “fabulous,” but nevertheless delayed sentencing and set a hearing on counsel’s motions for the following month.

During the motion hearing, Weeden’s trial attorney and Lisa Perrine, Ph.D., a forensic psychologist, testified. While on the stand, Weeden’s trial attorney, who was a former assistant district attorney that had litigated roughly 100 cases during his 20-year career, including at least 10 murder trials, stated that he considered retaining a forensic psychologist, but opted not to because he believed such testimony “wouldn’t necessarily have been helpful.” Additionally, he indicated that he believed such testimony could actually be harmful: “if the doctor were to conclude that [Weeden] was completely immature for instance and – and easy – easily manipulated, my concern would have been that the Prosecution would have used that to show that she didn’t understand perhaps the magnitude of a robbery, but still participated in it.” Weeden’s trial attorney also noted that he believed the witnesses testifying about Weeden’s “involvement and what was said prior to the robbery . . . lacked credibility.” Dr. Perrine, who met with Weeden for a total of 7 hours following her conviction, testified that she believed that Weeden “probably had the understanding of what a robbery was when she was age 14,” but, because Weeden “misses really important information in her environment,” she “may not put all the pieces together and understand that a robbery is intended.” When initially asked if Weeden “ha[d] enough information that she was capable of processing, to share the intent with the perpetrator that the victim be robbed,” Dr. Perrine answered, “I don’t know.” However, after Weeden’s new counsel told her that the co-defendants were older than Weeden, Dr. Perrine seemingly changed her position and indicated that Weeden would not necessarily share the co- defendants’ intent.

On cross examination, Dr. Perrine admitted that, besides speaking with Weeden, Weeden’s parents, and Weeden’s teacher, the only evidence she reviewed in forming her opinions was the presentence report. Thus, Dr. Perrine did not read the police reports in the case, the witnesses’ statements to law enforcement, the phone records, the text messages exchanged between Weeden and Moore, or the transcript from trial. According to Dr. Perrine, “even if [she] had reviewed all of that information, [she was] not at all sure that it would have made a whole lot of difference.”

In addition to testimony, Weeden’s new counsel also produced a declaration from Weeden and an expert report from Dr. Perrine. In her declaration, Weeden stated, among other things, that she did not know that Moore and Melonson intended to rob the victims. As for Dr. Perrine’s report, despite the fact that Weeden had never been held back in her schooling, enrolled in a special education class, participated in outpatient psychotherapy, or reported any mental health

WEEDEN V. JOHNSON 27

symptoms, Dr. Perrine opined that Weeden “functions intellectually in the low average to borderline range with 91% of adolescents her age functioning at a higher level.” Dr. Perrine’s report concluded that Weeden could likely comprehend the concept of robbery and that someone else intended to commit one, but that “it [wa]s extremely unlikely she would intend to commit robbery or knowingly participate in one.”

After considering the evidence in support of Weeden’s motion, as well as the parties’ briefing and oral arguments, the Superior Court denied Weeden’s claim that her trial attorney was ineffective. It did so based on its finding that Weeden had failed to satisfy both prongs of the Strickland v. Washington, 466 U.S. 668 (1984), test. Specifically, the Superior Court found that Weeden’s trial attorney’s decision not to employ a forensic psychologist was “a sound tactical decision” because Weeden’s “intent d[id] not appear to be at issue since she den[ied] that she planned the robbery or agreed to it or had any conversation about it when she set the victim up.” The Superior Court also added that “it d[id] not appear [that trial counsel’s decision] ha[d] been prejudicial in any event.”

On direct appeal, Weeden’s new counsel again claimed that his predecessor was constitutionally ineffective by “failing to present psychological testimony” that Weeden was “[in]capable of forming the intent to rob.” People v. Melonson, 2013 WL 1987240, at *20 (Cal. Ct. App. May 15, 2013), as modified on denial of reh’g (June 14, 2013). The California Court of Appeal rejected this claim, stating:

Weeden contends trial counsel’s decision not to have her examined by a psychologist was not a sound tactical decision. However, trial counsel articulated a very reasonable tactical decision for not requesting such an examination. If the examination revealed Weeden was easily manipulated, the prosecution could claim this was evidence that although Weeden did not understand the magnitude of the robbery, she nonetheless went along with it. In essence, the examination might undermine the defense strategy trial counsel was pursuing at trial.

In addition, Weeden argues the lack of psychological testimony on her lack of specific intent “undermines confidence in the outcome of this trial.” According to Weeden, the testimony of Dr. Perrine would have raised a reasonable doubt as to whether Weeden harbored a specific intent to rob. Presented with Dr. Perrine’s testimony, the jury would not have concluded Weeden joined in Melonson’s criminal intent to commit armed robbery.

Again, trial counsel pursued the defense that Weeden did not set up the robbery or agree to it. The testimony Weeden claims counsel was ineffective in failing to present might well have undermined this defense. Among Dr. Perrine’s findings was that Weeden had a strong tendency to be a passive follower, which might have led the jury to conclude she did not initiate the robbery but,

WEEDEN V. JOHNSON 29

sheep-like, went along with it. We find no ineffective assistance.

Id. at *21.

The California Supreme Court summarily denied Weeden’s petition for review. The District Court also denied Weeden’s habeas petition. See Weeden v. Johnson, 2014 WL 6808778 (E.D. Cal. Nov. 26, 2014).

II. DISCUSSION

A federal habeas petition challenging state custody shall be denied “unless the state court’s adjudication of the claim[] resulted in a decision that was contrary to, or involved an unreasonable application of, clearly established Federal law, as determined by the Supreme Court of the United States.” 28 U.S.C. § 2254(d)(1) (emphasis added). “A decision is ‘contrary to’ Supreme Court precedent where ‘the state court arrives at a conclusion opposite to that reached by the Supreme Court on a question of law or if the state court decides a case differently than the Supreme Court has on a set of materially indistinguishable facts.’” Jones v. Harrington, 829 F.3d 1128, 1135 (9th Cir. 2016) (alterations in original omitted) (quoting Williams v. Taylor, 529 U.S. 362, 413 (2000)). A state court unreasonably applies clearly established federal law if it “identifies the correct governing legal rule but unreasonably applies it to the facts of the particular state prisoner’s case.” White v. Woodall, 134 S. Ct. 1697, 1705 (2014) (alteration in original and internal quotation marks omitted). State court decisions are to be measured “against [the Supreme Court’s] precedents as of the time the state court renders its decision.” Greene v. Fisher, 132 S. Ct. 38, 44 (2011) (emphasis in original) (internal quotation marks omitted).

Here, the majority concludes that the California Court of Appeal’s finding that Weeden’s trial counsel’s performance was adequate is “contrary to, or involved an unreasonable application of, clearly established Supreme Court law.” Maj. Op. 14–15 (internal quotation marks omitted). To be constitutionally ineffective, an attorney’s performance must not only be deficient, but the deficiency must also be prejudicial. See Strickland, 466 U.S. at 687. “Surmounting Strickland’s high bar is never an easy task[,]” Padilla v. Kentucky, 559 U.S. 356, 371 (2010), and is made “doubly” hard when a federal court is reviewing a state court’s application of Strickland, Knowles v. Mirzayance, 556 U.S. 111, 123 (2009). Here, neither the record nor existing case law support granting habeas relief on ineffective assistance grounds.2

A. The California Court of Appeal’s determination that Weeden’s trial attorney’s performance was adequate is not contrary to nor an unreasonable application of clearly established Supreme Court law.

According to the majority, Weeden’s trial attorney was constitutionally ineffective by failing to have a psychological examination performed on Weeden because “Weeden’s mental condition was an essential factor in deciding whether she actually had the required mental states for the [charged]

2 Weeden also seeks habeas relief based on the Superior Court’s refusal to instruct the jury on Weeden’s age. Although not reached by the majority, I would deny relief on that ground as well.

WEEDEN V. JOHNSON 31

crime.” Maj. Op. 13–14 (internal quotation marks omitted). But what Supreme Court precedent mandates this conclusion? While the majority is correct that an attorney has a “duty to make reasonable investigations or to make a reasonable decision that makes particular investigations unnecessary,” Strickland, 466 U.S. at 691, we are not to frame the issue presented at a “high level of generality.” Lopez v. Smith, 135 S. Ct. 1, 4 (2014) (per curiam) (internal quotation marks omitted). Rather, when deciding if a decision is contrary to established law, we must ask if a Supreme Court case addresses “the specific question presented by this case,” Woods v. Donald, 135 S. Ct. 1372, 1377 (2015) (per curiam), which here is whether a defendant must be psychologically examined for the guilt phase of a non-capital case in which competency is not an issue. No Supreme Court case has answered this question in the affirmative. Therefore, the California Court of Appeal’s determination that Weeden’s trial counsel’s decision not to request an examination was a very reasonable tactical decision “[can]not be contrary to any holding from [the Supreme Court].” Id. (internal quotation marks omitted).

Additionally, the California Court of Appeal’s determination is not an unreasonable application of Supreme Court precedent. “The Strickland standard is a general one, so the range of reasonable applications is substantial.” Harrington v. Richter, 562 U.S. 86, 105 (2011). Further, under AEDPA, “the question is not whether counsel’s actions were reasonable,” but, rather, “whether there is any reasonable argument that counsel satisfied Strickland’s deferential standard.” Id. at 105. In other words, “[t]he state court decision must be so lacking in justification that there was an error well understood and comprehended in existing law beyond any possibility for fairminded disagreement.” Woods v. Etherton, 136 S. Ct. 1149, 1151 (2016) (per curiam).

At a minimum, there could be fairminded disagreement about whether Weeden’s trial attorney’s representation was adequate. An attorney provides effective assistance if their representation “does not fall below an objective standard of reasonableness in light of prevailing professional norms.” Bobby v. Van Hook, 558 U.S. 4, 7 (2009) (per curiam). Prevailing professional norms are typically distilled from standards developed by the American Bar Association (“ABA”). See Wiggins v. Smith, 539 U.S. 510, 524 (2003) (noting that the Supreme Court has long referred to ABA standards as “guides to determining what is reasonable” (internal quotation marks omitted)). The Criminal Justice Standards for the Defense Function published by the ABA do not mention, much less require, psychological evaluations in non-capital cases, which is unlike what the ABA has said in the capital context. See ABA’s Guidelines for the Appointment and Performance of Defense Counsel in Death Penalty Cases, Guideline 4.1.A.1-2 (rev. ed. 2003) (stating that a capital-defense team “should contain at least one member qualified by training and experience to screen individuals for the presence of mental or psychological disorders or impairments”). Thus, the California Court of Appeal’s determination that Weeden’s trial attorney was effective is not inconsistent with any then-existing professional norm.

What’s more, a fairminded judge, considering all of the circumstances in this case, could believe that Weeden’s trial attorney’s decision not to have Weeden examined was sound. When trial began, Weeden’s trial attorney had represented Weeden for over two and a half years. Thus, the attorney had

WEEDEN V. JOHNSON 33

sufficient time to familiarize himself with the case and his client and to formulate a defense strategy. Further, after litigating approximately 100 cases, including at least 10 murder trials, Weeden’s trial attorney had the experience to make sound defense strategies. Last, Weeden’s trial attorney arguably executed his strategy at trial effectively because Weeden was acquitted of a serious offense—attempted murder—and the Superior Court described the trial attorney’s performance as “excellent” and “fabulous.” As a result, reasonable minds could well believe that Weeden’s trial attorney was effective.

The majority’s conclusion to the contrary is based on an incomplete reading of Harrington v. Richter, 562 U.S. 86 (2011). First, while the Harrington court did state that “[c]riminal cases will arise where the only reasonable and available defense strategy requires consultation with experts,” it qualified this observation by indicating that such cases are “[r]are.” 562 U.S. at 106. This case is not one of those rare cases where the evidence against the defendant was so overwhelming or pointed so strongly in one direction that the only reasonable defense available was one that incorporated psychological evidence. Weeden was not at the scene, there was no physical evidence directly tying Weeden to the charged crimes, all of the co-defendants were older, and Weeden had no criminal record. Plus, the People’s case heavily relied upon statements made by co-defendants or relatives of the co-defendants, all of whom had made inconsistent statements. Thus, each statement made by the People’s key witnesses was challengeable on the ground that its declarant was biased and/or untrustworthy. Accordingly, this case is one where a reasonable attorney could choose to base his defense on the theory that the People’s evidence was insufficient to meet reasonable doubt’s stringent requirement, as opposed to grasping for straws and claiming that his competent client somehow lacked the ability to form the requisite intent. See id. at 109 (“To support a defense argument that the prosecution has not proved its case it sometimes is better to try to cast pervasive suspicion of doubt than to strive to prove a certainty that exonerates.”).

Second, the majority’s reading of Harrington overlooks the case’s holding that an attorney does not have to consult with every expert “whose insight might possibly have been useful.” Id. at 107. “An attorney can avoid activities that appear distractive from more important duties.” Id. (internal quotation marks omitted). Further, an attorney’s “attention to certain issues to the exclusion of others” is “strong[ly] presum[ed]” to reflect a “trial tactic[] rather than sheer neglect.” Id. at 109 (internal quotation marks omitted).

Here, Weeden’s trial attorney could have reasonably believed that his time and available resources were better spent on the defense theory that Weeden was unaware that Melonson and Moore intended to rob the victims than on the theory that she was incapable of intending for someone to be robbed. The former theory was not a clear loser, as evidenced by the jury’s partial acquittal. While the majority may have taken a different approach, “[r]eliance on the harsh light of hindsight to cast doubt on a trial that took place now more than [8] years ago is precisely what Strickland and AEDPA seek to prevent.” Id. (internal quotation marks omitted).

At bottom, the majority’s determination that Weeden is entitled to habeas relief can only be justified if there is a per se rule that a defense attorney must seek a psychological evaluation whenever his or her client arguably has some

WEEDEN V. JOHNSON 35

mental impairment, regardless of whether the attorney believes such an evaluation will be valuable to the defense. However, no Supreme Court precedent establishes such a per se rule. In fact, such precedent indicates that there are few, if any, per se rules in the ineffective assistance counsel context: “[r]are are the situations in which the wide latitude counsel must have in making tactical decisions will be limited in any one technique or approach.” Harrington, 562 U.S. at 106 (internal quotation marks omitted). As a result, the majority errs in holding that California Court of Appeal acted contrary to or unreasonably applied established Supreme Court precedent by finding Weeden’s trial attorney’s performance adequate.

B. Even if Weeden’s trial attorney’s performance was constitutionally deficient, Weeden was not prejudiced.

Realizing that it cannot clear the high hurdle AEDPA sets for the Strickland prejudice prong, the majority attempts an end-run around it by claiming that AEDPA deference does not apply in this case because the California Court of Appeal did not address prejudice. Maj. Op. 15. However, this is simply not true. Far from pretermitting the analysis, the California Court of Appeal identified a specific harm that could have occurred had Weeden’s trial attorney introduced testimony from a forensic psychologist: “Dr. Perrine’s finding[] was that Weeden had a strong tendency to be a passive follower, which might have led the jury to conclude she did not initiate the robbery but, sheep-like, went along with it.” Melonson, 2013 WL 1987240, at *21. While the majority may disagree with the California Court of Appeal’s assessment, this disagreement does not change the fact that the state court did address the issue and that its determination is entitled to deference.

As for the majority’s claim that the California Court of Appeal applied the wrong legal standard, Maj. Op. 16–18, it too does not hold water. Leading up to its prejudice analysis, the Court of Appeal correctly identified the governing legal standard—reasonable probability. See Melonson, 2013 WL 1987240, at *21 (stating that the defendant must show that “it is reasonably probable that a more favorable result would have been reached absent the deficient performance”). Furthermore, the Court of Appeal recited Weeden’s argument for why the reasonable probability standard was met in her case—the lack of psychological evidence “undermine[d] confidence in the outcome of th[e] trial.” Id. Against this backdrop, the Court of Appeal’s observation that Weeden’s psychological evidence “might have” actually hurt her case had it been offered does not indicate that it applied the wrong legal standard. While the majority may prefer that the Court of Appeal had offered a more complete written analysis for its no-prejudice finding, the fact that it did not does not mean that its finding should receive any less deference. See Johnson v. Williams, 133 S. Ct. 1088, 1095–96 (2013) (“The caseloads shouldered by many state appellate courts are very heavy, and the opinions issued by these courts must be read with that factor in mind.” (internal footnote omitted)).

Nonetheless, even if the majority were correct that we are to apply de novo review to the prejudice prong in this case, habeas relief should still be denied. Relief is warranted only if the trial attorney’s errors are “so serious as to deprive the defendant of a fair trial, a trial whose result is reliable.” Harrington, 562 U.S. at 104. “The likelihood of a different result must be substantial, not just conceivable.” Id. at 112.

WEEDEN V. JOHNSON 37

To say the least, the utility of calling a forensic psychologist to discuss Weeden’s age, maturity, or intelligence is questionable. To begin with, the psychological evidence offered by Weeden’s new counsel was weak. As noted in Dr. Perrine’s report, Weeden had never been held back in her schooling, enrolled in a special education class, participated in outpatient psychotherapy, or reported any mental health symptoms. Furthermore, Dr. Perrine’s opinions were based on incomplete information. For instance, Dr. Perrine had not reviewed the texts exchanged between Weeden and Moore following the shooting, which demonstrated that Weeden was no shrinking violet and was capable of trying to direct others’ activities—e.g., advising Moore that “[i]f [he] want[ed] to get out of this, [he would] have to give up Angie’s name.” Also, Dr. Perrine was unaware of Hill’s alleged warning to Weeden just prior to the shooting not to participate in the robbery because it could go bad.

Further diminishing the value of psychological evidence in this case is the fact that, even without it, the jury already had the information necessary to determine, if it wanted to, that Weeden lacked the requisite mental condition to intend a robbery. The jury was aware of Weeden’s age at the time of the shooting and was instructed that it could use its “common sense and experience” when evaluating the evidence. Thus, this case is fundamentally different from those where the Supreme Court has found that an attorney’s failure to investigate prejudiced the defendant. See, e.g., Wiggins, 539 U.S. at 534–35 (finding that the capital-murder defendant had been prejudiced during sentencing by his attorney’s failure to uncover evidence that the defendant had “experienced severe privation and abuse in the first six years of his life” and had been “physical[ly] torment[ed], sexual[ly] molest[ed], and repeated[ly] rape[d] during his subsequent years in foster care”).

In addition to psychological evidence adding little value to the defense, it might well have been harmful. As noted by the California Court of Appeal, a psychologist’s testimony would have been susceptible to being usurped by the prosecutor, as it could be manipulated to establish that Weeden’s age made her more likely to engage in or go along with risky or illegal behavior. Additionally, it likely would have invited the Government to call its own forensic psychologist and “transform the case into a battle of the experts.” Harrington, 562 U.S. at 108–09. In light of the text messages between Weeden and Moore, as well as the fact that Weeden had never been held back in her schooling or reported any mental health symptoms, shifting the jury’s focus from whether Weeden intended for the victims to be robbed to whether she had ability to intend a robbery likely would not have been in Weeden’s best interest.

Conclusion

The majority’s apparent desire to save Sarah Weeden from the consequences of her felony murder conviction is understandable—the result it reaches under AEDPA is not. While the majority may disagree with California’s decision to allow for the adult prosecution of 14-year-olds, nothing prohibits the State from adopting such a policy. What Weeden was entitled to under the law—a fair trial and competent counsel—she received. To be sure, it would be nice to magically undo Weeden’s decision to lure two unsuspecting victims into what proved to be a death trap. Unfortunately, we are federal judges, not magicians, and no do-over of the night in question is forthcoming. As a result,

WEEDEN V. JOHNSON 39

what we are stuck with is a conviction that has been reviewed and upheld by the California Court of Appeal. The state court’s determination is neither contrary to nor an unreasonable application of Supreme Court precedent, and, thus, there is no basis to afford Weeden habeas relief. Because the majority opinion reaches the opposite conclusion, I respectfully dissent.