Weaver v. Massachusetts

Justice KENNEDY delivered the opinion of the Court. During petitioner's trial on state criminal charges, the courtroom was occupied by potential jurors and closed to the public for two days of the jury selection process. Defense counsel neither objected to the closure at trial nor raised the issue on direct review. And the case comes to the Court on the assumption that, in failing to object, defense counsel provided ineffective assistance. In the direct review context, the underlying constitutional violation-the courtroom closure-has been treated by this Court as a structural error, i.e., an error entitling the defendant to automatic reversal without any inquiry into prejudice. The question is whether invalidation of the conviction is required here as well, or if the prejudice inquiry is altered when the structural error is raised in the context of an ineffective-assistance-of-counsel claim. I In 2003, a 15-year-old boy was shot and killed in Boston. A witness saw a young man fleeing the scene of the crime and saw him pull out a pistol. A baseball hat fell off of his head. The police recovered the hat, which featured a distinctive airbrushed Detroit Tigers logo on either side. The hat's distinctive markings linked it to 16-year-old Kentel Weaver. He is the petitioner here. DNA obtained from the hat matched petitioner's DNA. Two weeks after the crime, the police went to petitioner's house to question him. He admitted losing his hat around the time of the shooting but denied being involved. Petitioner's mother was not so sure. Later, she questioned petitioner herself. She asked whether he had been at the scene of the shooting, and he said he had been there. But when she asked if he was the shooter, or if he knew who the shooter was, petitioner put his head down and said nothing. Believing his response to be an admission of guilt, she insisted that petitioner go to the police station to confess. He did. Petitioner was indicted in Massachusetts state court for first-degree murder and the unlicensed possession of a handgun. He pleaded not guilty and proceeded to trial. *1906 The pool of potential jury members was large, some 60 to 100 people. The assigned courtroom could accommodate only 50 or 60 in the courtroom seating. As a result, the trial judge brought all potential jurors into the courtroom so that he could introduce the case and ask certain preliminary questions of the entire venire panel. Many of the potential jurors did not have seats and had to stand in the courtroom. After the preliminary questions, the potential jurors who had been standing were moved outside the courtroom to wait during the individual questioning of the other potential jurors. The judge acknowledged that the hallway was not "the most comfortable place to wait" and thanked the potential jurors for their patience. 2 Tr. II-103 (Apr. 10, 2006). The judge noted that there was simply not space in the courtroom for everybody. As all of the seats in the courtroom were occupied by the venire panel, an officer of the court excluded from the courtroom any member of the public who was not a potential juror. So when petitioner's mother and her minister came to the courtroom to observe the two days of jury selection, they were turned away. All this occurred before the Court's decision in Presley v. Georgia, 558 U.S. 209 , 130 S.Ct. 721 , 175 L.Ed.2d 675 (2010) ( per curiam ). Presley made it clear that the public-trial right extends to jury selection as well as to other portions of the trial. Id., at 213-215 , 130 S.Ct. 721 . Before Presley, Massachusetts courts would often close courtrooms to the public during jury selection, in particular during murder trials. In this case petitioner's mother told defense counsel about the closure at some point during jury selection. But counsel "believed that a courtroom closure for [jury selection] was constitutional." Crim. No. 2003-11293 (Super. Ct. Mass., Feb. 22, 2013), App. to Pet. for Cert. 49a. As a result, he "did not discuss the matter" with petitioner, or tell him "that his right to a public trial included the [jury voir dire ]," or object to the closure. Ibid . During the ensuing trial, the government presented strong evidence of petitioner's guilt. Its case consisted of the incriminating details outlined above, including petitioner's confession to the police. The jury convicted petitioner on both counts. The court sentenced him to life in prison on the murder charge and to about a year in prison on the gun-possession charge. Five years later, petitioner filed a motion for a new trial in Massachusetts state court. As relevant here, he argued that his attorney had provided ineffective assistance by failing to object to the courtroom closure. After an evidentiary hearing, the trial court recognized a violation of the right to a public trial based on the following findings: The courtroom had been closed; the closure was neither de minimis nor trivial; the closure was unjustified; and the closure was full rather than partial (meaning that all members of the public, rather than only some of them, had been excluded from the courtroom). The trial court further determined that defense counsel failed to object because of "serious incompetency, inefficiency, or inattention." Id., at 63a (quoting Massachusetts v. Chleikh, 82 Mass.App. 718 , 722, 978 N.E.2d 96 , 100 (2012) ). On the other hand, petitioner had not "offered any evidence or legal argument establishing prejudice." App. to Pet. for Cert. 64a. For that reason, the court held that petitioner was not entitled to relief. Petitioner appealed the denial of the motion for a new trial to the Massachusetts Supreme Judicial Court. The court consolidated that appeal with petitioner's *1907 direct appeal. As noted, there had been no objection to the closure at trial; and the issue was not raised in the direct appeal. The Supreme Judicial Court then affirmed in relevant part. Although it recognized that "[a] violation of the Sixth Amendment right to a public trial constitutes structural error," the court stated that petitioner had "failed to show that trial counsel's conduct caused prejudice warranting a new trial." 474 Mass. 787 , 814, 54 N.E.3d 495 , 520 (2016). On this reasoning, the court rejected petitioner's claim of ineffective assistance of counsel. There is disagreement among the Federal Courts of Appeals and some state courts of last resort about whether a defendant must demonstrate prejudice in a case like this one-in which a structural error is neither preserved nor raised on direct review but is raised later via a claim alleging ineffective assistance of counsel. Some courts have held that, when a defendant shows that his attorney unreasonably failed to object to a structural error, the defendant is entitled to a new trial without further inquiry. See, e.g., Johnson v. Sherry, 586 F.3d 439 , 447 (C.A.6 2009) ; Owens v. United States, 483 F.3d 48 , 64-65 (C.A.1 2007) ; Littlejohn v. United States, 73 A.3d 1034 , 1043-1044 (D.C.2013) ; State v. Lamere, 327 Mont. 115 , 125, 112 P.3d 1005 , 1013 (2005). Other courts have held that the defendant is entitled to relief only if he or she can show prejudice. See, e.g., Purvis v. Crosby, 451 F.3d 734 , 738 (C.A.11 2006) ; United States v. Gomez, 705 F.3d 68 , 79-80 (C.A.2 2013) ; Reid v. State, 286 Ga. 484 , 487, 690 S.E.2d 177 , 180-181 (2010). This Court granted certiorari to resolve that disagreement. 580 U.S. ----, 137 S.Ct. 809 , 196 L.Ed.2d 595 (2017). The Court does so specifically and only in the context of trial counsel's failure to object to the closure of the courtroom during jury selection. II This case requires a discussion, and the proper application, of two doctrines: structural error and ineffective assistance of counsel. The two doctrines are intertwined; for the reasons an error is deemed structural may influence the proper standard used to evaluate an ineffective-assistance claim premised on the failure to object to that error. A The concept of structural error can be discussed first. In Chapman v. California, 386 U.S. 18 , 87 S.Ct. 824 , 17 L.Ed.2d 705 (1967), this Court "adopted the general rule that a constitutional error does not automatically require reversal of a conviction." Arizona v. Fulminante, 499 U.S. 279 , 306, 111 S.Ct. 1246 , 113 L.Ed.2d 302 (1991) (citing Chapman,supra ). If the government can show "beyond a reasonable doubt that the error complained of did not contribute to the verdict obtained," the Court held, then the error is deemed harmless and the defendant is not entitled to reversal. Id., at 24, 87 S.Ct. 824 . The Court recognized, however, that some errors should not be deemed harmless beyond a reasonable doubt. Id., at 23, n. 8 , 87 S.Ct. 824 . These errors came to be known as structural errors. See Fulminante, 499 U.S., at 309-310 , 111 S.Ct. 1246 . The purpose of the structural error doctrine is to ensure insistence on certain basic, constitutional guarantees that should define the framework of any criminal trial. Thus, the defining feature of a structural error is that it "affect[s] the framework within which the trial proceeds," rather than being "simply an error in the trial process itself." Id ., at 310, 111 S.Ct. 1246 . For the same reason, a structural error "def[ies] analysis by harmless *1908 error standards." Id., at 309 , 111 S.Ct. 1246 (internal quotation marks omitted). The precise reason why a particular error is not amenable to that kind of analysis-and thus the precise reason why the Court has deemed it structural-varies in a significant way from error to error. There appear to be at least three broad rationales. First, an error has been deemed structural in some instances if the right at issue is not designed to protect the defendant from erroneous conviction but instead protects some other interest. This is true of the defendant's right to conduct his own defense, which, when exercised, "usually increases the likelihood of a trial outcome unfavorable to the defendant." McKaskle v. Wiggins, 465 U.S. 168 , 177, n. 8, 104 S.Ct. 944 , 79 L.Ed.2d 122 (1984). That right is based on the fundamental legal principle that a defendant must be allowed to make his own choices about the proper way to protect his own liberty. See Faretta v. California, 422 U.S. 806 , 834, 95 S.Ct. 2525 , 45 L.Ed.2d 562 (1975). Because harm is irrelevant to the basis underlying the right, the Court has deemed a violation of that right structural error. See United States v. Gonzalez-Lopez, 548 U.S. 140 , 149, n. 4, 126 S.Ct. 2557 , 165 L.Ed.2d 409 (2006). Second, an error has been deemed structural if the effects of the error are simply too hard to measure. For example, when a defendant is denied the right to select his or her own attorney, the precise "effect of the violation cannot be ascertained." Ibid. (quoting Vasquez v. Hillery, 474 U.S. 254 , 263, 106 S.Ct. 617 , 88 L.Ed.2d 598 (1986) ). Because the government will, as a result, find it almost impossible to show that the error was "harmless beyond a reasonable doubt," Chapman, supra, at 24 , 87 S.Ct. 824 , the efficiency costs of letting the government try to make the showing are unjustified. Third, an error has been deemed structural if the error always results in fundamental unfairness. For example, if an indigent defendant is denied an attorney or if the judge fails to give a reasonable-doubt instruction, the resulting trial is always a fundamentally unfair one. See Gideon v. Wainwright, 372 U.S. 335 , 343-345, 83 S.Ct. 792 , 9 L.Ed.2d 799 (1963) (right to an attorney); Sullivan v. Louisiana, 508 U.S. 275 , 279, 113 S.Ct. 2078 , 124 L.Ed.2d 182 (1993) (right to a reasonable-doubt instruction). It therefore would be futile for the government to try to show harmlessness. These categories are not rigid. In a particular case, more than one of these rationales may be part of the explanation for why an error is deemed to be structural. See e.g., id., at 280-282 , 113 S.Ct. 2078 . For these purposes, however, one point is critical: An error can count as structural even if the error does not lead to fundamental unfairness in every case. See Gonzalez-Lopez, supra, at 149 , n. 4, 126 S.Ct. 2557 (rejecting as "inconsistent with the reasoning of our precedents" the idea that structural errors "always or necessarily render a trial fundamentally unfair and unreliable" (emphasis deleted)). B As noted above, a violation of the right to a public trial is a structural error. See supra, at 1905, 1906 - 1907. It is relevant to determine why that is so. In particular, the question is whether a public-trial violation counts as structural because it always leads to fundamental unfairness or for some other reason. In Waller v. Georgia, 467 U.S. 39 , 104 S.Ct. 2210 , 81 L.Ed.2d 31 (1984), the state court prohibited the public from viewing a *1909 weeklong suppression hearing out of concern for the privacy of persons other than those on trial. See id., at 41-43 , 104 S.Ct. 2210 . Although it recognized that there would be instances where closure was justified, this Court noted that "such circumstances will be rare" and that the closure in question was unjustified. Id., at 45, 48 , 104 S.Ct. 2210 . Still, the Court did not order a new trial. Id., at 49-50 , 104 S.Ct. 2210 . Instead it ordered a new suppression hearing that was open to the public. Id., at 50 , 104 S.Ct. 2210 . If the same evidence was found admissible in that renewed pretrial proceeding, the Court held, no new trial as to guilt would be necessary. Ibid. This was despite the structural aspect of the violation. Some 25 years after the Waller decision, the Court issued its per curiam ruling in Presley v. Georgia . 558 U.S. 209 , 130 S.Ct. 721 , 175 L.Ed.2d 675 . In that case, as here, the courtroom was closed to the public during jury voir dire . Id., at 210 , 130 S.Ct. 721 . Unlike here, however, there was a trial objection to the closure, and the issue was raised on direct appeal. Id., at 210-211 , 130 S.Ct. 721 . On review of the State Supreme Court's decision allowing the closure, this Court expressed concern that the state court's reasoning would allow the courtroom to be closed during jury selection "whenever the trial judge decides, for whatever reason, that he or she would prefer to fill the courtroom with potential jurors rather than spectators." Id ., at 215, 130 S.Ct. 721 (internal quotation marks omitted). Although the Court expressly noted that courtroom closure may be ordered in some circumstances, the Court also stated that it was "still incumbent upon" the trial court "to consider all reasonable alternatives to closure." Id., at 215-216 , 130 S.Ct. 721 . These opinions teach that courtroom closure is to be avoided, but that there are some circumstances when it is justified. The problems that may be encountered by trial courts in deciding whether some closures are necessary, or even in deciding which members of the public should be admitted when seats are scarce, are difficult ones. For example, there are often preliminary instructions that a judge may want to give to the venire as a whole, rather than repeating those instructions (perhaps with unintentional differences) to several groups of potential jurors. On the other hand, various constituencies of the public-the family of the accused, the family of the victim, members of the press, and other persons-all have their own interests in observing the selection of jurors. How best to manage these problems is not a topic discussed at length in any decision or commentary the Court has found. So although the public-trial right is structural, it is subject to exceptions. See Simonson, The Criminal Court Audience in a Post-Trial World, 127 Harv. L. Rev. 2173 , 2219-2222 (2014) (discussing situations in which a trial court may order a courtroom closure). Though these cases should be rare, a judge may deprive a defendant of his right to an open courtroom by making proper factual findings in support of the decision to do so. See Waller, supra, at 45 , 104 S.Ct. 2210 . The fact that the public-trial right is subject to these exceptions suggests that not every public-trial violation results in fundamental unfairness. A public-trial violation can occur, moreover, as it did in Presley, simply because the trial court omits to make the proper findings before closing the courtroom, even if those findings might have been fully supported by the evidence. See 558 U.S., at 215 , 130 S.Ct. 721 . It would be unconvincing to deem a trial fundamentally unfair just because a judge omitted to announce factual findings before making *1910 an otherwise valid decision to order the courtroom temporarily closed. As a result, it would be likewise unconvincing if the Court had said that a public-trial violation always leads to a fundamentally unfair trial. Indeed, the Court has not said that a public-trial violation renders a trial fundamentally unfair in every case. In the two cases in which the Court has discussed the reasons for classifying a public-trial violation as structural error, the Court has said that a public-trial violation is structural for a different reason: because of the "difficulty of assessing the effect of the error." Gonzalez-Lopez, 548 U.S., at 149, n. 4 , 126 S.Ct. 2557 ; see also Waller, supra, at 49, n. 9 , 104 S.Ct. 2210 . The public-trial right also protects some interests that do not belong to the defendant. After all, the right to an open courtroom protects the rights of the public at large, and the press, as well as the rights of the accused. See, e.g., Press-Enterprise Co. v. Superior Court of Cal., Riverside Cty., 464 U.S. 501 , 508-510, 104 S.Ct. 819 , 78 L.Ed.2d 629 (1984) ; Richmond Newspapers, Inc. v. Virginia, 448 U.S. 555 , 572-573, 100 S.Ct. 2814 , 65 L.Ed.2d 973 (1980). So one other factor leading to the classification of structural error is that the public-trial right furthers interests other than protecting the defendant against unjust conviction. These precepts confirm the conclusion the Court now reaches that, while the public-trial right is important for fundamental reasons, in some cases an unlawful closure might take place and yet the trial still will be fundamentally fair from the defendant's standpoint. III The Court now turns to the proper remedy for addressing the violation of a structural right, and in particular the right to a public trial. Despite its name, the term "structural error" carries with it no talismanic significance as a doctrinal matter. It means only that the government is not entitled to deprive the defendant of a new trial by showing that the error was "harmless beyond a reasonable doubt." Chapman, 386 U.S., at 24 , 87 S.Ct. 824 . Thus, in the case of a structural error where there is an objection at trial and the issue is raised on direct appeal, the defendant generally is entitled to "automatic reversal" regardless of the error's actual "effect on the outcome." Neder v. United States, 527 U.S. 1 , 7, 119 S.Ct. 1827 , 144 L.Ed.2d 35 (1999). The question then becomes what showing is necessary when the defendant does not preserve a structural error on direct review but raises it later in the context of an ineffective-assistance-of-counsel claim. To obtain relief on the basis of ineffective assistance of counsel, the defendant as a general rule bears the burden to meet two standards. First, the defendant must show deficient performance-that the attorney's error was "so serious that counsel was not functioning as the 'counsel' guaranteed the defendant by the Sixth Amendment." Strickland v. Washington, 466 U.S. 668 , 687, 104 S.Ct. 2052 , 80 L.Ed.2d 674 (1984). Second, the defendant must show that the attorney's error "prejudiced the defense." Ibid. The prejudice showing is in most cases a necessary part of a Strickland claim. The reason is that a defendant has a right to effective representation, not a right to an attorney who performs his duties "mistake-free." Gonzalez-Lopez, 548 U.S., at 147 , 126 S.Ct. 2557 . As a rule, therefore, a "violation of the Sixth Amendment right to effective representation is not 'complete' until the defendant is prejudiced." Ibid. *1911 (emphasis deleted); see also Premo v. Moore, 562 U.S. 115 , 128, 131 S.Ct. 733 , 178 L.Ed.2d 649 (2011) ; Lockhart v. Fretwell, 506 U.S. 364 , 370, 113 S.Ct. 838 , 122 L.Ed.2d 180 (1993). That said, the concept of prejudice is defined in different ways depending on the context in which it appears. In the ordinary Strickland case, prejudice means "a reasonable probability that, but for counsel's unprofessional errors, the result of the proceeding would have been different." 466 U.S., at 694 , 104 S.Ct. 2052 . But the Strickland Court cautioned that the prejudice inquiry is not meant to be applied in a "mechanical" fashion. Id., at 696 , 104 S.Ct. 2052 . For when a court is evaluating an ineffective-assistance claim, the ultimate inquiry must concentrate on "the fundamental fairness of the proceeding." Ibid . Petitioner therefore argues that under a proper interpretation of Strickland, even if there is no showing of a reasonable probability of a different outcome, relief still must be granted if the convicted person shows that attorney errors rendered the trial fundamentally unfair. For the analytical purposes of this case, the Court will assume that petitioner's interpretation of Strickland is the correct one. In light of the Court's ultimate holding, however, the Court need not decide that question here. As explained above, not every public-trial violation will in fact lead to a fundamentally unfair trial. See supra, at 1910. Nor can it be said that the failure to object to a public-trial violation always deprives the defendant of a reasonable probability of a different outcome. Thus, when a defendant raises a public-trial violation via an ineffective-assistance-of-counsel claim, Strickland prejudice is not shown automatically. Instead, the burden is on the defendant to show either a reasonable probability of a different outcome in his or her case or, as the Court has assumed for these purposes, see supra, at 1910 - 1911, to show that the particular public-trial violation was so serious as to render his or her trial fundamentally unfair. Neither the reasoning nor the holding here calls into question the Court's precedents determining that certain errors are deemed structural and require reversal because they cause fundamental unfairness, either to the defendant in the specific case or by pervasive undermining of the systemic requirements of a fair and open judicial process. See Murray, A Contextual Approach to Harmless Error Review, 130 Harv. L. Rev. 1791 , 1813, 1822 (2017) (noting that the "eclectic normative objectives of criminal procedure" go beyond protecting a defendant from erroneous conviction and include ensuring " 'that the administration of justice should reasonably appear to be disinterested' " (quoting Liljeberg v. Health Services Acquisition Corp., 486 U.S. 847 , 869-870, 108 S.Ct. 2194 , 100 L.Ed.2d 855 (1988) )). Those precedents include Sullivan v. Louisiana, 508 U.S., at 278-279 , 113 S.Ct. 2078 (failure to give a reasonable-doubt instruction); Tumey v. Ohio, 273 U.S. 510 , 535, 47 S.Ct. 437 , 71 L.Ed. 749 (1927) (biased judge); and Vasquez v. Hillery, 474 U.S., at 261-264 , 106 S.Ct. 617 (exclusion of grand jurors on the basis of race). See Neder, supra, at 8 , 119 S.Ct. 1827 (describing each of these errors as structural). This Court, in addition, has granted automatic relief to defendants who prevailed on claims alleging race or gender discrimination in the selection of the petit jury, see Batson v. Kentucky, 476 U.S. 79 , 100, 106 S.Ct. 1712 , 90 L.Ed.2d 69 (1986) ; J.E.B. v. Alabama ex rel. T. B., 511 U.S. 127 , 145-146, 114 S.Ct. 1419 , 128 L.Ed.2d 89 (1994), though the Court has yet to label those errors structural in express terms, see, e.g., Neder, supra, at 8 , 119 S.Ct. 1827 . The errors in those cases necessitated *1912 automatic reversal after they were preserved and then raised on direct appeal. And this opinion does not address whether the result should be any different if the errors were raised instead in an ineffective-assistance claim on collateral review. The reason for placing the burden on the petitioner in this case, however, derives both from the nature of the error, see supra, at 1910 - 1912, and the difference between a public-trial violation preserved and then raised on direct review and a public-trial violation raised as an ineffective-assistance-of-counsel claim. As explained above, when a defendant objects to a courtroom closure, the trial court can either order the courtroom opened or explain the reasons for keeping it closed. See supra , at 1909 - 1910. When a defendant first raises the closure in an ineffective-assistance claim, however, the trial court is deprived of the chance to cure the violation either by opening the courtroom or by explaining the reasons for closure. Furthermore, when state or federal courts adjudicate errors objected to during trial and then raised on direct review, the systemic costs of remedying the error are diminished to some extent. That is because, if a new trial is ordered on direct review, there may be a reasonable chance that not too much time will have elapsed for witness memories still to be accurate and physical evidence not to be lost. There are also advantages of direct judicial supervision. Reviewing courts, in the regular course of the appellate process, can give instruction to the trial courts in a familiar context that allows for elaboration of the relevant principles based on review of an adequate record. For instance, in this case, the factors and circumstances that might justify a temporary closure are best considered in the regular appellate process and not in the context of a later proceeding, with its added time delays. When an ineffective-assistance-of-counsel claim is raised in postconviction proceedings, the costs and uncertainties of a new trial are greater because more time will have elapsed in most cases. The finality interest is more at risk, see Strickland, 466 U.S., at 693-694 , 104 S.Ct. 2052 (noting the "profound importance of finality in criminal proceedings"), and direct review often has given at least one opportunity for an appellate review of trial proceedings. These differences justify a different standard for evaluating a structural error depending on whether it is raised on direct review or raised instead in a claim alleging ineffective assistance of counsel. In sum, "[a]n ineffective-assistance claim can function as a way to escape rules of waiver and forfeiture and raise issues not presented at trial," thus undermining the finality of jury verdicts. Harrington v. Richter, 562 U.S. 86 , 105, 131 S.Ct. 770 , 178 L.Ed.2d 624 (2011). For this reason, the rules governing ineffective-assistance claims "must be applied with scrupulous care." Premo, 562 U.S., at 122 , 131 S.Ct. 733 . IV The final inquiry concerns the ineffective-assistance claim in this case. Although the case comes on the assumption that petitioner has shown deficient performance by counsel, he has not shown prejudice in the ordinary sense, i.e., a reasonable probability that the jury would not have convicted him if his attorney had objected to the closure. It is of course possible that potential jurors might have behaved differently if petitioner's family had been present. And it is true that the presence of the public might have had some bearing on juror reaction. But here petitioner offered no *1913 "evidence or legal argument establishing prejudice" in the sense of a reasonable probability of a different outcome but for counsel's failure to object. App. to Pet. for Cert. 64a; see Strickland, 466 U.S., at 694 , 104 S.Ct. 2052 . In other circumstances a different result might obtain. If, for instance, defense counsel errs in failing to object when the government's main witness testifies in secret, then the defendant might be able to show prejudice with little more detail. See ibid. Even in those circumstances, however, the burden would remain on the defendant to make the prejudice showing, id., at 694, 696 , 104 S.Ct. 2052 , because a public-trial violation does not always lead to a fundamentally unfair trial, see supra, at 1910. In light of the above assumption that prejudice can be shown by a demonstration of fundamental unfairness, see supra, at 1910 - 1911, the remaining question is whether petitioner has shown that counsel's failure to object rendered the trial fundamentally unfair. See Strickland, supra, at 696 , 104 S.Ct. 2052 The Court concludes that petitioner has not made the showing. Although petitioner's mother and her minister were indeed excluded from the courtroom for two days during jury selection, petitioner's trial was not conducted in secret or in a remote place. Cf. In re Oliver, 333 U.S. 257 , 269, n. 22, 68 S.Ct. 499 , 92 L.Ed. 682 (1948). The closure was limited to the jury voir dire ; the courtroom remained open during the evidentiary phase of the trial; the closure decision apparently was made by court officers rather than the judge; there were many members of the venire who did not become jurors but who did observe the proceedings; and there was a record made of the proceedings that does not indicate any basis for concern, other than the closure itself. There has been no showing, furthermore, that the potential harms flowing from a courtroom closure came to pass in this case. For example, there is no suggestion that any juror lied during voir dire ; no suggestion of misbehavior by the prosecutor, judge, or any other party; and no suggestion that any of the participants failed to approach their duties with the neutrality and serious purpose that our system demands. It is true that this case comes here on the assumption that the closure was a Sixth Amendment violation. And it must be recognized that open trials ensure respect for the justice system and allow the press and the public to judge the proceedings that occur in our Nation's courts. Even so, the violation here did not pervade the whole trial or lead to basic unfairness. In sum, petitioner has not shown a reasonable probability of a different outcome but for counsel's failure to object, and he has not shown that counsel's shortcomings led to a fundamentally unfair trial. He is not entitled to a new trial. * * * In the criminal justice system, the constant, indeed unending, duty of the judiciary is to seek and to find the proper balance between the necessity for fair and just trials and the importance of finality of judgments. When a structural error is preserved and raised on direct review, the balance is in the defendant's favor, and a new trial generally will be granted as a matter of right. When a structural error is raised in the context of an ineffective-assistance claim, however, finality concerns are far more pronounced. For this reason, and in light of the other circumstances present in this case, petitioner must show prejudice in order to obtain a new trial. As explained above, he has not made the required showing. The judgment of the *1914 Massachusetts Supreme Judicial Court is affirmed. It is so ordered. Justice THOMAS, with whom Justice GORSUCH joins, concurring. I write separately with two observations about the scope of the Court's holding. First, this case comes to us on the parties' "assumption[s]" that the closure of the courtroom during jury selection "was a Sixth Amendment violation" and that "defense counsel provided ineffective assistance" by "failing to object" to it. Ante, at 1905, 1913. The Court previously held in a per curiam opinion-issued without the benefit of merits briefing or argument-that the Sixth Amendment right to a public trial extends to jury selection. See Presley v. Georgia, 558 U.S. 209 , 213, 130 S.Ct. 721 , 175 L.Ed.2d 675 (2010) ; id., at 216 , 130 S.Ct. 721 (THOMAS, J., dissenting). I have some doubts about whether that holding is consistent with the original understanding of the right to a public trial, and I would be open to reconsidering it in a case in which we are asked to do so. Second, the Court "assume[s]," for the "analytical purposes of this case," that a defendant may establish prejudice under Strickland v. Washington, 466 U.S. 668 , 104 S.Ct. 2052 , 80 L.Ed.2d 674 (1984), by demonstrating that his attorney's error led to a fundamentally unfair trial. Ante, at 1910 - 1911. According to Strickland , a defendant may establish prejudice by showing "a reasonable probability that, but for counsel's unprofessional errors, the result of the proceeding would have been different"; by showing an "[a]ctual or constructive denial of the assistance of counsel altogether"; or by showing that counsel labored under "an actual conflict of interest." 466 U.S., at 692-694 , 104 S.Ct. 2052 . Strickland did not hold, as the Court assumes, that a defendant may establish prejudice by showing that his counsel's errors "rendered the trial fundamentally unfair." Ante, at 1911. Because the Court concludes that the closure during petitioner's jury selection did not lead to fundamental unfairness in any event, ante, at 1913 - 1914, no part of the discussion about fundamental unfairness, see ante, at 1910 - 1913, is necessary to its result. In light of these observations, I do not read the opinion of the Court to preclude the approach set forth in Justice ALITO's opinion, which correctly applies our precedents. Justice ALITO, with whom Justice GORSUCH joins, concurring in the judgment. This case calls for a straightforward application of the familiar standard for evaluating ineffective assistance of counsel claims. Strickland v. Washington, 466 U.S. 668 , 687, 104 S.Ct. 2052 , 80 L.Ed.2d 674 (1984). Weaver cannot meet that standard, and therefore his claim must be rejected. The Sixth Amendment protects a criminal defendant's right "to have the Assistance of Counsel for his defence." That right is violated when (1) "counsel's performance was deficient" in the relevant sense of the term and (2) "the deficient performance prejudiced the defense." Strickland, supra, at 687 , 104 S.Ct. 2052 The prejudice requirement-which is the one at issue in this case-"arises from the very nature" of the right to effective representation: Counsel simply "cannot be 'ineffective' unless his mistakes have harmed the defense (or, at least, unless it is reasonably likely that they have)." United States v. Gonzalez-Lopez, 548 U.S. 140 , 147, 126 S.Ct. 2557 , 165 L.Ed.2d 409 (2006). In other words, "a violation of the Sixth Amendment right to effective representation *1915 is not 'complete' until the defendant is prejudiced." Ibid. Strickland 's definition of prejudice is based on the reliability of the underlying proceeding. "The benchmark for judging any claim of ineffectiveness must be whether counsel's conduct so undermined the proper functioning of the adversarial process that the trial cannot be relied on as having produced a just result." 466 U.S., at 686 , 104 S.Ct. 2052 (emphasis added); see United States v. Cronic, 466 U.S. 648 , 658, 104 S.Ct. 2039 , 80 L.Ed.2d 657 (1984). This is so because "[t]he purpose of the Sixth Amendment guarantee of counsel is to ensure that a defendant has the assistance necessary to justify reliance on the outcome of the proceeding." Strickland, 466 U.S., at 691-692 , 104 S.Ct. 2052 . Accordingly, an attorney's error "does not warrant setting aside the judgment of a criminal proceeding if the error had no effect on the judgment." Id., at 691 , 104 S.Ct. 2052 . Weaver makes much of the Strickland Court's statement that "the ultimate focus of inquiry must be on the fundamental fairness of the proceeding." Id., at 696 , 104 S.Ct. 2052 . But the very next sentence clarifies what the Court had in mind, namely, the reliability of the proceeding. In that sentence, the Court explains that the proper concern-"[i]n every case"-is "whether, despite the strong presumption of reliability, the result of the particular proceeding is unreliable." Ibid. In other words, the focus on reliability is consistent throughout the Strickland opinion. To show that a counsel's error rendered a legal proceeding unreliable, a defendant ordinarily must demonstrate "a reasonable probability that, but for counsel's unprofessional errors, the result of the proceeding would have been different." Id., at 694 , 104 S.Ct. 2052 . In a challenge to a conviction, such as the one in this case, this means that the defendant must show "a reasonable probability that, absent the errors, the factfinder would have had a reasonable doubt respecting guilt." Id., at 695 , 104 S.Ct. 2052 . The Court has relieved defendants of the obligation to make this affirmative showing in only a very narrow set of cases in which the accused has effectively been denied counsel altogether: These include the actual or constructive denial of counsel, state interference with counsel's assistance, or counsel that labors under actual conflicts of interest. Id., at 692 , 104 S.Ct. 2052 ; Cronic, 466 U.S., at 658-660 , 104 S.Ct. 2039 . Prejudice can be presumed with respect to these errors because they are "so likely to prejudice the accused that the cost of litigating their effect in a particular case is unjustified." Id., at 658 , 104 S.Ct. 2039 ; see Strickland, supra, at 692 , 104 S.Ct. 2052 ; Mickens v. Taylor, 535 U.S. 162 , 175, 122 S.Ct. 1237 , 152 L.Ed.2d 291 (2002). In short, there are two ways of meeting the Strickland prejudice requirement. A defendant must demonstrate either that the error at issue was prejudicial or that it belongs to the narrow class of attorney errors that are tantamount to a denial of counsel, for which an individualized showing of prejudice is unnecessary. Weaver attempts to escape this framework by stressing that the deprivation of the right to a public trial has been described as a "structural" error, but this is irrelevant under Strickland . The concept of "structural error" comes into play when it is established that an error occurred at the trial level and it must be decided whether the error was harmless. See Neder v . United States, 527 U.S. 1 , 7, 119 S.Ct. 1827 , 144 L.Ed.2d 35 (1999) ; Arizona v . Fulminante, 499 U.S. 279 , 309-310, 111 S.Ct. 1246 , 113 L.Ed.2d 302 (1991). The prejudice prong of Strickland *1916 is entirely different. It does not ask whether an error was harmless but whether there was an error at all, for unless counsel's deficient performance prejudiced the defense, there was no Sixth Amendment violation in the first place. See Gonzalez-Lopez, supra, at 150 , 126 S.Ct. 2557 (even where an attorney's deficient performance "pervades the entire trial," "we do not allow reversal of a conviction for that reason without a showing of prejudice" because "the requirement of showing prejudice in ineffectiveness claims stems from the very definition of the right at issue"). Weaver's theory conflicts with Strickland because it implies that an attorney's error can be prejudicial even if it "had no effect," or only "some conceivable effect," on the outcome of his trial. Strickland, supra, at 691, 693, 104 S.Ct. 2052 . That is precisely what Strickland rules out. To sum up, in order to obtain relief under Strickland, Weaver must show that the result of his trial was unreliable. He could do so by demonstrating a reasonable likelihood that his counsel's error affected the verdict. Alternatively, he could establish that the error falls within the very short list of errors for which prejudice is presumed. Weaver has not attempted to make either argument, so his claim must be rejected. I would affirm the judgment of the Supreme Judicial Court of Massachusetts on that ground. Justice BREYER, with whom Justice KAGAN joins, dissenting. The Court notes that Strickland' s "prejudice inquiry is not meant to be applied in a 'mechanical' fashion," ante, at 1911 (quoting Strickland v. Washington, 466 U.S. 668 , 696, 104 S.Ct. 2052 , 80 L.Ed.2d 674 (1984) ), and I agree. But, in my view, it follows from this principle that a defendant who shows that his attorney's constitutionally deficient performance produced a structural error should not face the additional-and often insurmountable- Strickland hurdle of demonstrating that the error changed the outcome of his proceeding. In its harmless-error cases, this Court has "divided constitutional errors into two classes": trial errors and structural errors. United States v. Gonzalez-Lopez, 548 U.S. 140 , 148, 126 S.Ct. 2557 , 165 L.Ed.2d 409 (2006). Trial errors are discrete mistakes that "occu[r] during the presentation of the case to the jury." Arizona v. Fulminante, 499 U.S. 279 , 307, 111 S.Ct. 1246 , 113 L.Ed.2d 302 (1991). Structural errors, on the other hand, "affec[t] the framework within which the trial proceeds." Id., at 310 , 111 S.Ct. 1246 . The Court has recognized that structural errors' distinctive attributes make them "defy analysis by 'harmless-error' standards." Id., at 309 , 111 S.Ct. 1246 . It has therefore categorically exempted structural errors from the case-by-case harmlessness review to which trial errors are subjected. Our precedent does not try to parse which structural errors are the truly egregious ones. It simply views all structural errors as "intrinsically harmful" and holds that any structural error warrants "automatic reversal" on direct appeal "without regard to [its] effect on the outcome" of a trial. Neder v. United States, 527 U.S. 1 , 7, 119 S.Ct. 1827 , 144 L.Ed.2d 35 (1999). The majority here does not take this approach. It assumes that some structural errors-those that "lead to fundamental unfairness"-but not others, can warrant relief without a showing of actual prejudice under Strickland . Ante, at 1908 - 1909, 1910 - 1912. While I agree that a showing of fundamental unfairness is sufficient to satisfy Strickland, I would not try to draw this distinction. *1917 Even if some structural errors do not create fundamental unfairness, all structural errors nonetheless have features that make them "defy analysis by 'harmless-error' standards." Fulminante, supra, at 309 , 111 S.Ct. 1246 . This is why all structural errors-not just the "fundamental unfairness" ones-are exempt from harmlessness inquiry and warrant automatic reversal on direct review. Those same features mean that all structural errors defy an actual-prejudice analysis under Strickland . For instance, the majority concludes that some errors-such as the public-trial error at issue in this case-have been labeled "structural" because they have effects that "are simply too hard to measure." Ante, at 1908; see, e.g., Sullivan v. Louisiana, 508 U.S. 275 , 281-282, 113 S.Ct. 2078 , 124 L.Ed.2d 182 (1993) (explaining that structural errors have "consequences that are necessarily unquantifiable and indeterminate"). But how could any error whose effects are inherently indeterminate prove susceptible to actual-prejudice analysis under Strickland ? Just as the "difficulty of assessing the effect" of such an error would turn harmless-error analysis into "a speculative inquiry into what might have occurred in an alternate universe," Gonzalez-Lopez, supra, at 149, n. 4, 150, 126 S.Ct. 2557 , so too would it undermine a defendant's ability to make an actual-prejudice showing to establish an ineffective-assistance claim. The problem is evident with regard to public-trial violations. This Court has recognized that "the benefits of a public trial are frequently intangible, difficult to prove, or a matter of chance." Waller v. Georgia, 467 U.S. 39 , 49, n. 9, 104 S.Ct. 2210 , 81 L.Ed.2d 31 (1984). As a result, "a requirement that prejudice be shown 'would in most cases deprive [the defendant] of the [public-trial] guarantee, for it would be difficult to envisage a case in which he would have evidence available of specific injury.' " Ibid. (quoting United States ex rel. Bennett v. Rundle, 419 F.2d 599 , 608 (C.A.3 1969) (en banc)) (alteration in original). In order to establish actual prejudice from an attorney's failure to object to a public-trial violation, a defendant would face the nearly impossible burden of establishing how his trial might have gone differently had it been open to the public. See ibid. (" '[D]emonstration of prejudice in this kind of case is a practical impossibility ...' " (quoting State v. Sheppard, 182 Conn. 412 , 418, 438 A.2d 125 , 128 (1980) )). I do not see how we can read Strickland as requiring defendants to prove what this Court has held cannot be proved. If courts do not presume prejudice when counsel's deficient performance leads to a structural error, then defendants may well be unable to obtain relief for incompetence that deprived them "of basic protections without which a criminal trial cannot reliably serve its function as a vehicle for determination of guilt or innocence." Neder, supra, at 8-9 , 119 S.Ct. 1827 (internal quotation marks omitted). This would be precisely the sort of "mechanical" application that Strickland tells us to avoid. In my view, we should not require defendants to take on a task that is normally impossible to perform. Nor would I give lower courts the unenviably complex job of deciphering which structural errors really undermine fundamental fairness and which do not-that game is not worth the candle. I would simply say that just as structural errors are categorically insusceptible to harmless-error analysis on direct review, so too are they categorically insusceptible to actual-prejudice analysis in Strickland claims. A showing that an attorney's constitutionally deficient performance produced a structural error should consequently *1918 be enough to entitle a defendant to relief. I respectfully dissent.