dissenting from the denial of rehearing en banc.
The Supreme Court’s jurisprudence makes clear that imposing a visible restraint on the accused is inherently prejudicial to his right to a fair trial. John M. Stephenson was required to wear a stun belt during the guilt phase of his capital trial without any inquiry by the judge as to whether such a restraint was justified nor any record that might have supported such a finding. Although the stun belt was no doubt intended to be a discreet restraint, the belt underneath Stephenson’s clothes was, in fact, readily visible to the jury and anyone else in his purview. My colleagues have concluded that Stephenson has not shown that he was harmed by his attorney’s unexplained failure to object to the stun belt and thus cannot demonstrate prejudice for purposes of his claim that he was deprived of the effective assistance of counsel. See Strickland v. Washington, 466 U.S. 668, 691-96, 104 S.Ct. 2052, 2066-69, 80 L.Ed.2d 674 (1984). I believe their analysis overlooks the inherent, unquantifiable prejudice of a visible restraint and is otherwise inconsistent with the Supreme Court’s decisions on this subject.
Physically restraining the accused in front of the jury has long been regarded as an unavoidably prejudicial act that is justified only in extraordinary circumstances. See Illinois v. Allen, 397 U.S. 337, 344, 90 S.Ct. 1057, 1061, 25 L.Ed.2d 353 (1970); Holbrook v. Flynn, 475 U.S. 560, 568, 106 S.Ct. 1340, 1345, 89 L.Ed.2d 525 (1986); *733Deck v. Missouri, 544 U.S. 622, 626-29, 125 S.Ct. 2007, 2010-12, 161 L.Ed.2d 953 (2005). A visible restraint bespeaks a belief on the part of the judiciary that the defendant must be physically separated from the community at large, including the jury. Id. at 630, 125 S.Ct. at 2013 (quoting Holbrook, 475 U.S. at 569, 106 S.Ct. at 1346). It communicates to the jury that the defendant is untrustworthy, out of control, likely to flee, menacing, bad. See Roche v. Davis, 291 F.3d 473, 482-83 (7th Cir.2002) (“the sight of a defendant in shackles ‘could instill in the jury a belief that the defendant is a dangerous individual who cannot be controlled, an idea that could be devastating to his defense’”) (quoting Harrell v. Israel, 672 F.2d 632, 637 (7th Cir.1982) (per curiam)). As such, the restraint undermines the presumption that the defendant is innocent of the crime charged and poses an affront to the dignity of the courtroom proceeding. Deck, 544 U.S. at 680-32, 125 S.Ct. at 2013; Holbrook, 475 U.S. at 568, 106 S.Ct. at 1345; Allen, 397 U.S. at 344, 90 S.Ct. at 1061. To the extent it hobbles the defendant’s movements (or, in the case of a stun belt, preoccupies his mind with worry that he may be zapped with a 50,000-volt jolt of electricity), it may also interfere with the defendant’s ability to participate in his own defense. Deck, 544 U.S. at 631, 125 S.Ct. at 2013; Wrinkles v. State, 749 N.E.2d 1179, 1194-95 (Ind.2001). A visible restraint is, consequently, “a last resort,” Allen, 397 U.S. at 344, 90 S.Ct. at 1061, that is consistent with due process only when the trial judge determines that the restraint is justified by an essential state interest specific to a particular trial. Deck, 544 U.S. at 629, 632, 125 S.Ct. at 2012, 2014; see also Holbrook, 475 U.S. at 568-69, 106 S.Ct. at 1345-46; Allen, 397 U.S. at 344, 90 S.Ct. at 1061.
Although the Supreme Court’s cases on restraints have dealt with shackles rather than stun belts, there is no reason to think that the Court would treat a visible stun belt any differently from other types of visible restraints. However technologically removed a stun belt may be from its clanking predecessors, it serves the same incapacitating function and, if perceived by the jury, will be understood as the modern-day equivalent of manacles. A stun belt that is visible to the jury will thus communicate all of the same negative messages about the need to physically control the defendant and separate him from the community. This court and others have therefore presumed that a visible stun belt is as prejudicial to the defendant’s due process rights as other visible restraints. See Wrinkles v. Buss, 537 F.3d 804, 814-15 (7th Cir.2008) (majority); id. at 830-31 (dissent); United States v. Miller, 531 F.3d 340, 344-45 (6th Cir.2008); Gonzalez v. Pliler, 341 F.3d 897, 900 (9th Cir.2003); United States v. Durham, 287 F.3d 1297, 1306 (11th Cir.2002); see also id. at 1305 (visible stun belt “ ‘may be even more prejudicial than handcuffs or leg irons because it implies that unique force is necessary to control the defendant’ ”) (quoting State v. Flieger, 91 Wash.App. 236, 955 P.2d 872, 874 (1998)); Gonzalez, 341 F.3d at 900-01 (noting that stun belts may have more of a psychological impact on defendant than shackles and therefore may be more likely to interfere with his ability to participate in his defense). More to the point, the Indiana Supreme Court itself has deemed a stun belt to be equally prejudicial to the defendant. Stephenson v. State, 864 N.E.2d 1022, 1033 (Ind.2007) (“The use of a stun belt, if perceived by the jury, produces all of the results that shackling does.”).1 We must defer to that conclusion *734so long as it represents a reasonable application of the Supreme Court’s precedents on physical restraints, as it surely does. 28 U.S.C. § 2254(d)(1); see, e.g., Waddington v. Sarausad, 555 U.S. 179, 129 S.Ct. 823, 831, 172 L.Ed.2d 532 (2009).
Beginning with jury selection, Stephenson was required to wear a stun belt during both the guilt and penalty phases of his trial for murder in Indiana state court, without any finding that he posed a risk of escape, violence, or disruptive behavior that might justify such a physical restraint. Indeed, so far as the record reveals, no one thought that there was a need to restrain Stephenson during the trial. Rather, as a matter of general security protocol, the sheriff wanted Stephenson restrained during his transport between the courthouse and the jail. During pretrial proceedings, this was accomplished by means of wrist and ankle restraints along with a “belly chain” connected to the former. But as the trial approached, the sheriff decided that a stun belt worn underneath Stephenson’s shirt would serve the same purpose, while avoiding the possibility that jurors might catch a glimpse from the jury room window of Stephenson emerging from or entering the transport vehicle in shackles. Postconviction Tr. 532, 538. All well and good. What remains a mystery is why anyone thought the stun belt needed to remain on Stephenson’s person once he had been safely delivered to the courthouse. Neither the sheriff nor any of the four other officers in charge of security at Stephenson’s trial believed that he posed a security threat or had engaged in any behavior during the six-month period of his pretrial confinement that might warrant continued restraint within the courtroom. Postconviction Tr. 525, 532-33, 540, 544, 548. On the contrary, Stephenson had turned himself in when he learned that he was wanted by the authorities, he had made no attempts at escape, and, as both the Indiana Supreme Court and this court noted, he had comported himself as a “gentleman” and “a model prisoner” from the time of his arrest forward. 864 N.E.2d at 1036-37; Stephenson v. Wilson, 619 F.3d 664, 667 (7th Cir.2010). Nonetheless, the belt was kept on his person once he was delivered to court.
For his part, Stephenson’s attorney (who himself had no concerns about Stephenson’s behavior, see Postconviction Tr. 100-01) never challenged the premise that Stephenson required some sort of restraint even while in the courtroom and instead acceded to the sheriffs decision without ever bringing the issue before the trial judge. Counsel testified in the postconviction proceeding that, in his experience, the trial judge typically deferred to the sheriffs security decisions. Postconviction Tr. 149-50. “I understood our choices were either that [the stun belt] or shaekles[,] and that [i.e., shackling] was certainly not an acceptable alternative.” Postconviction Tr. 102. As a result, the trial judge never considered whether Stephenson ought to be restrained by means of the stun belt or any other form of restraint during the trial proceedings.
To the extent that the judge, the sheriff, and the parties assumed that the belt was not visible to the jury, they were wrong. Although Stephenson wore the belt underneath his clothing, it created a visible bulge underneath the back of his shirt. A videotape of local news coverage of the trial, admitted into evidence at the post-conviction hearing, shows Stephenson being led from the courthouse to a waiting vehicle. Petitioner’s Postconviction Ex. 45. I have collected several screenshots from that video in an appendix to this dissent. When Stephenson is observed *735from either the side or the rear, a bulge the size of a large fanny pack is clearly visible beneath his long-sleeve dress shirt. There is no dispute that, as a result of that bulge, some of the jurors became aware during the trial that Stephenson was wearing a stun belt. In the postconviction proceeding, Stephenson produced affidavits and deposition testimony from four jurors to that effect.2 Based on that evidence, the Indiana Supreme Court specifically found that “Stephenson has established by a preponderance of the evidence that the belt was ‘readily visible’ to the jury.” 864 N.E.2d at 1034. Stephenson was thus exposed to all of the prejudicial consequences of visible restraints, without any case-specific evidence, let alone findings by the trial judge, that might justify such an exceptional measure. This was a clear violation of his right, under the Fifth and Fourteenth Amendments, to due process. See Deck, 544 U.S. at 629, 125 S.Ct. at 2012.
Had Stephenson’s counsel preserved the issue by objecting to the stun belt at trial, the burden would have fallen to the State to prove beyond a reasonable doubt that the belt did not contribute to the jury’s verdict. Id. at 635, 125 S.Ct. at 2015-16. Deck reiterates the Court’s prior observation in Holbrook that visible restraints are “ ‘inherently prejudicial’ ” to the defendant, such that, in a head-on challenge to such restraints, he need not submit proof establishing that the restraints harmed him in some demonstrable way. Id., 125 S.Ct. at 2015 (quoting Holbrook, 475 U.S. at 568, 106 S.Ct. at 1345).
That statement is rooted in our belief that the practice will often have negative effects, but — like “the consequences of compelling a defendant to wear prison clothing” or of forcing him to stand trial medicated' — those effects “cannot be shown from a trial transcript.” Riggins [v. Nevada], supra, [504 U.S. 127,] at 137, 112 S.Ct. 1810 [at 1816, 118 L.Ed.2d 479 (1992) ]. Thus, where a court, without adequate justification, orders the defendant to wear shackles that will be seen by the jury, the defendant need not demonstrate actual prejudice to make out a due process violation....
Deck, 544 U.S. at 635, 125 S.Ct. at 2015.
But because Stephenson’s trial counsel never challenged the need for a restraint, and thus did not preserve a freestanding due process claim, Stephenson has instead pursued a claim that his trial counsel’s failure to object to the stun belt deprived him of the effective assistance of counsel to which he was entitled under the Sixth and Fourteenth Amendments. See Murray v. Carrier, 477 U.S. 478, 496, 106 S.Ct. 2639, *7362649-50, 91 L.Ed.2d 397 (1986). Under the familiar framework of Strickland, 466 U.S. at 687, 104 S.Ct. at 2064, Stephenson must establish both that his attorney’s performance fell below objective norms and that he was prejudiced by the unprofessional conduct.
There is no dispute that Stephenson has met the first of these two requirements. As the Indiana Supreme Court noted, “at the time of Stephenson’s trial it was well-settled as a matter of both state law and the requirements of federal due process that no form of visible restraint was permissible without an individualized finding that the defendant presented a risk of escape, violence, or disruption of the trial.” 864 N.E.2d at 1032 (citing Deck, 544 U.S. at 626-27, 125 S.Ct. at 2011, and Coates v. State, 487 N.E.2d 167, 169 (Ind.Ct.App.1985)). Thus, when the sheriff decided that Stephenson would be restrained while in the courtroom, without any indication that such a restraint was necessary, it was counsel’s obligation to demand a hearing as to the need for such a restraint. Instead, counsel acquiesced in the presumption that some form of restraint was warranted, content that a stun belt would be used in lieu of shackles because a stun belt was less likely to be noticed by the jury.
The Indiana Supreme Court did not quarrel with counsel’s preference for a stun belt over shackles, but it did find fault with “[cjounsel’s unquestioning acceptance of the need for any form of restraint.” Id. Given the absence of evidence that Stephenson posed any risk of flight or disruption to the court proceeding, there was no legitimate tactical reason not to object to a restraint and to insist on a judicial finding as to the need for such a restraint. Id. at 1032, 1035. “On this record, ... failure to object to the belt cannot be justified as a tactical decision ... because the explanation offered by counsel for their decision boiled down to a failure to know the applicable law.” Id. at 1035. Counsel’s failure to object to the stun belt thus fell below the objective standard of reasonable representation, and in this respect deprived Stephenson of the effective representation to which the Constitution entitled him. This was a reasonable application of Strickland as well as the Supreme Court’s precedents on visible restraints, and it commands our deference. § 2254(d)(1); Waddington, 555 U.S. 179, 129 S.Ct. at 831, 172 L.Ed.2d 532.
The remaining question is whether Stephenson was prejudiced by his counsel’s ineffectiveness, and on this point the Indiana Supreme Court’s truncated analysis is indefensible. Notwithstanding the court’s conclusion that counsel was obliged to object to the stun belt given the lack of evidence warranting a restraint, the court went on to reason that the failure to pose such an objection caused Stephenson no harm because the trial judge inevitably would have overruled such an objection. 864 N.E.2d at 1040-41. The court noted that Stephenson was accused of committing a triple murder that had the hallmarks of a premeditated assassination related to narcotics trafficking. Id. Moreover, according to the witness who was present when Stephenson committed the murders, Stephenson had threatened to kill the witness if he told anyone what Stephenson had done. Id. at 1041. “[Gjiven the state of the law in 1996, we think it plain that the trial judge would have followed the sheriffs recommendation and ordered that the belt be deployed at the guilt phase even if defendant’s counsel had objected and required a hearing and findings as to the need for its use.” Id.
The flaw in this reasoning is patent: the same lack of evidence that obliged counsel to object to a restraint would have compelled the trial judge to sustain counsel’s *737objection. In applying Strickland’s prejudice prong to an attorney’s failure to lodge an objection, we must presume that the trial judge would have correctly applied the law in ruling on that objection. 466 U.S. at 694-95, 104 S.Ct. at 2068. “The assessment of prejudice should proceed on the assumption that the decisionmaker is reasonably, conscientiously, and impartially applying the standards that govern the decision.” Id. at 695, 104 S.Ct. at 2068. The trial judge thus could have overruled an objection and ordered that Stephenson be restrained only if there were evidence particular to Stephenson’s trial supporting such a measure. See Deck, 544 U.S. at 626-29, 125 S.Ct. at 2010-12 (summarizing well-established case law requiring such evidence). Although the trial judge had never had occasion to inquire into the need for a restraint (given the lack of an objection), that subject was aired in the post-conviction proceeding, where none of the five officials responsible for security at Stephenson’s trial articulated a basis for concern that Stephenson posed a risk of flight, disruption, or danger so as to warrant being restrained in the courtroom. Consequently, there was no particularized evidence of a need for restraint — none. In the Indiana Supreme Court’s own words, “the record shows nothing to support an individualized determination that Stephenson required any form of restraint at trial....” 864 N.E.2d at 1031 (emphasis supplied).
All that the Indiana Supreme Court could cite in support of the notion that some form of restraint was justified were the circumstances of the charged offense: the three murders, committed in a premeditated fashion, with a contemporaneous threat to a witness to remain silent or else. Of course, capital offenses by their very nature will nearly always involve the most violent and disturbing of criminal acts. If the nature of the offense were alone enough to support visible restraints, then such restraints would be routine in capital cases. In fact, as the panel pointed out, “[t]he cases ... hold that the nature of the crime with which a defendant is charged, however heinous, is insufficient by itself to justify visible restraints.” 619 F.3d at 668 (coll.cases); cf. Deck, 544 U.S. at 632-35, 125 S.Ct. at 2014-16 (finding restraints unjustified in penalty phase of prosecution for double murder of an elderly couple committed in course of robbery, even after finding of defendant’s guilt had been secured and affirmed on appeal). The state court was thus wrong, and plainly so, in reasoning that because the trial judge would have overruled an objection to the stun belt (or other visible restraint), Stephenson was not prejudiced by his attorney’s failure to object.
Given the lack of any evidence to support a visible restraint and the clear state of the law as to such restraints, only one ruling on an objection to the belt would have been proper, and that would be to sustain the objection. The Indian'a Supreme Court’s prejudice analysis was thus objectively unreasonable. See Roche v. Davis, supra, 291 F.3d at 483. That court never separately considered whether Stephenson was prejudiced in the sense that the visible stun belt might have affected the outcome of his trial. As there is no analysis from the state court on that point, our own analysis is necessarily plenary. Rompilla v. Beard, 545 U.S. 374, 390, 125 S.Ct. 2456, 2467, 162 L.Ed.2d 360 (2005) (citing Wiggins v. Smith, 539 U.S. 510, 534, 123 S.Ct. 2527, 2542, 156 L.Ed.2d 471 (2003)).
Stephenson must demonstrate a reasonable probability that, but for his counsel’s deficient performance, the outcome of the trial might have been different. Strickland, 466 U.S. at 694, 104 S.Ct. at 2068. Both the nature of the unobjected-to error and the weight of the evidence are key *738considerations in this inquiry. Id. at 695-96, 104 S.Ct. at 2069. Stephenson need not show that a different outcome was more likely than not; he need only show that the likelihood of a different outcome was better than negligible. See id. at 693, 104 S.Ct. at 2068; Gross v. Knight, 560 F.3d 668, 671 (7th Cir.), cert denied, — U.S. -, 130 S.Ct. 402, 175 L.Ed.2d 275 (2009).
Although the panel recognized that the evidence against Stephenson was not overwhelming, 619 F.3d at 673, it found no evidence demonstrating that the stun belt contributed to the jury’s finding of guilt. The panel noted that (1) the four juror affidavits simply demonstrated awareness of the stun belt, as opposed to any negative inferences the jurors might have drawn from the belt, and there are no reliable studies revealing the effects of such a restraint on a jury’s perception of the accused; (2) some security was required in a trial for a crime of violence, and had Stephenson not worn the stun belt, more guards would have been required in the courtroom, and it is unclear whether additional guards would have been less prejudicial to Stephenson in the jurors’ eyes than the visible bulge of the stun belt; (3) the trial lasted many months due to the vigorous efforts of the defense team, making it unlikely, in the panel’s view, that the jury would have given the stun belt much weight in assessing Stephenson’s guilt. Id. at 671-73.
I submit that the panel’s analysis starts off on the wrong foot in that it fails to acknowledge the inherent prejudice of a visible restraint. Yes, it is Stephenson who bears the burden under Strickland to establish a reasonable likelihood that, but for his attorney’s failure to object to the stun belt, he might have been acquitted. 466 U.S. at 687, 693-94, 104 S.Ct. at 2064, 2067-68. By contrast, had his attorney objected to the belt and had Stephenson mounted a freestanding challenge to the belt on review of his conviction, the burden instead would have fallen to the State to prove beyond a reasonable doubt that the belt had no effect on the conviction. Deck, 544 U.S. at 635, 125 S.Ct. at 2015-16. But simply because Strickland assigns the burden of persuasion to Stephenson does not cause the inherently prejudicial nature of a visible restraint to evaporate. Visible restraints have been deemed “a last resort” not because they are sometimes harmful to the defendant and sometimes not, depending on the circumstances of the particular case, but because they are always and unavoidably prejudicial to the defendant. That is why decisions like Deck and Holbrook describe them as inherently prejudicial. Id. at 635, 125 S.Ct. at 2015; Holbrook, 475 U.S. at 568, 106 S.Ct. at 1345. Deck further recognizes that this inherent prejudice is difficult if not impossible to document. 544 U.S. at 635, 125 S.Ct. at 2015. Thus, although Strickland puts Stephenson’s challenge to the stun belt in a different posture, it does not change the nature of the underlying error of requiring him to wear a visible restraint. Indeed, Strickland itself recognizes that nature of the error — for example, pervasive error versus isolated error — factors into the prejudice analysis. 466 U.S. at 695-96, 104 S.Ct. at 2069. Requiring a defendant to wear a stun belt without facts to support such a last-resort measure is a pervasive error, in that it affects the whole trial, the jury’s perception of the defendant, and such fundamental aspects of the prosecution as the presumption of innocence. The error might not ultimately prejudice the defendant in the sense that Strickland refers to prejudice — in other words, it might not alter the outcome of the trial, as when the proof of guilt is overwhelming (see, e.g., Roche, 291 F.3d at 484; Fountain v. United States, 211 F.3d 429, 436 (7th Cir. 2000)) — but it remains prejudicial in the *739sense that Deck and Holbrook discuss prejudice: it undermines the presumption of innocence, interferes with the defendant’s ability to participate in his own defense, and coarsens the courtroom environment. The iStrickland prejudice inquiry must begin with that recognition.
If one does begin the analysis by recognizing that a visible restraint is inherently prejudicial to the defendant, then it becomes clear that Stephenson has shown that the visible stun belt likely did affect the jury’s verdict as to his guilt. As the panel recognized, the evidence against Stephenson, although strong, was not overwhelming. 619 F.3d at 673. The Indiana Supreme Court conceded the same point, noting that its finding as to the lack of Strickland prejudice “does not rest on essentially indisputable evidence that establishes the merits of his conviction and sentence.” 864 N.E.2d at 1039 (emphasis supplied). There was no strong forensic evidence (fingerprints, DNA evidence, etc.) tying Stephenson to the murders. The State’s case rested primarily on the testimony of two witnesses, Dale Funk and Brian Mossberger, who themselves had connections to the crime: Funk was in Stephenson’s company before, during, and after the killings, and Mossberger was discovered to be in possession of the gun that was used to shoot the victims. At the same time, two other individuals, Jimmy Knight and Herschel Seifert, had made statements to (or within earshot of) third parties unconnected to Stephenson indicating that they were responsible for the murders. See Trial Tr. 29704, 30836-37, 31486, 31525-27, 31531. I do not doubt that the evidence against Stephenson was sufficient to convict him. But given the inherently prejudicial nature of a visible restraint and the lack of overwhelming evidence establishing Stephenson’s guilt, Stephenson has established a better than negligible probability that he might have been acquitted had he not been noticeably restrained, see Gross, 560 F.3d at 671, and this is “sufficient to undermine confidence in the outcome” of the guilt phase of the trial, Strickland, 466 U.S. at 694, 104 S.Ct. at 2068. This is exactly what Judge Springmann concluded in granting Stephenson relief. Stephenson v. Levenhagen, No. 3:07-CV-539-TS, 2009 WL 1886081, at *15 (N.D.Ill. Jul. 1, 2009). See Roche, 291 F.3d at 484 (finding that visible shackles prejudiced defendant in sentencing hearing, given that there was mitigating as well as aggravating evidence and in view of “the extreme inherent prejudice associated with shackling”).
The panel’s reasons for discounting the possibility that the jury’s verdict was affected by the stun belt are in tension with the Supreme Court’s visible-restraint jurisprudence. For example, the panel assumed that the likely alternative to the stun belt — more guards in the courtroom — might also have prejudiced Stephenson. 619 F.3d at 672. Yet, the Court in Holbrook expressly held that guards are not inherently prejudicial to the defendant in the way that visible restraints are: “While shackling and prison clothes are unmistakable indications of the need to separate a defendant from the community at large, the presence of guards at a defendant’s trial need not be interpreted as a sign that he is particularly dangerous or culpable.” 475 U.S. at 569, 106 S.Ct. at 1346.
Second, however vigorous a defense Stephenson’s counsel may have mounted on his behalf, it is impossible to say whether that defense compensated for a restraint that “undermines the presumption of innocence and the related fairness of the fact-finding process.” Deck, 544 U.S. at 630, 125 S.Ct. at 2013. To my mind, the fact that the evidence against Stephenson was not overwhelming, coupled with a vigorous defense case, simply makes it more rather *740than less likely that the pervasive and negative effects of a visible restraint may have been what tipped the jury toward a finding of guilt.
Finally, although the juror affidavits reveal nothing about the impact of the visible stun belt on the jury’s perceptions of Stephenson, such proof is unnecessary. “Whenever a courtroom arrangement is challenged as inherently prejudicial, ... the question must be not whether jurors actually articulated a consciousness of some prejudicial effect, but rather whether ‘an unacceptable risk is presented of impermissible factors coming into play.’ ” Holbrook, 475 U.S. at 570, 106 S.Ct. at 1346-47 (quoting Estelle v. Williams, 425 U.S. 501, 505, 96 S.Ct. 1691, 1693, 48 L.Ed.2d 126 (1976)). Indeed, the Court in Holbrook found that the lower court had properly disregarded the trial judge’s finding, based on the questionnaires completed by prospective jurors during voir dire, that the jury was likely unaffected by the presence of uniformed troopers at the defendant’s trial. “[Ljittle stock need be placed in jurors’ claims” that they would not be influenced by a practice challenged as inherently prejudicial, the Court observed. Id. at 570, 106 S.Ct. at 1346. (As noted, however, the Court went on to conclude that the presence of troopers was not, in fact, inherently prejudicial to the defendant as claimed.) Similarly, we cannot expect jurors who have already found the defendant guilty and recommended that he be sentenced to death to objectively assess whether and to what degree their deliberations were affected by the sight of the defendant visibly restrained.
The panel also remarked on the lack of careful empirical studies exploring the effects of visible stun belts and other restraints on jury deliberations. 619 F.3d at 673. Of course, Stephenson cannot be faulted for what social scientists have not yet documented. In the absence of such studies, we must “rely on our own experience and common sense,” Holbrook, 475 U.S. at 571 n. 4, 106 S.Ct. at 1347 n. 4, which tell us that a visible restraint conveys a distinctly negative message about the defendant to the jury. Any doubts on that score are resolved by the Supreme Court’s decisions recognizing the prejudice inherent in visible restraints.
Stephenson is entitled to a new trial, as Judge Springmann concluded. On this record, which reflects a total absence of case-specific evidence demonstrating a need for Stephenson to be restrained in the courtroom, there can be no dispute that being made to wear a readily-visible stun belt deprived Stephenson of due process. As the Indiana Supreme Court found, the failure of Stephenson’s trial counsel to object to the restraint was unsupported by any valid strategic reason and instead is explained solely by counsel’s ignorance of the case law prohibiting visible restraints except in extraordinary circumstances. A single failing by one’s counsel will support relief under Strickland when the error is both egregious and harmful to the defendant, Murray v. Carrier, supra, 477 U.S. at 496, 106 S.Ct. at 2649, as it was here. The inherently prejudicial nature of visible restraints, coupled with the lack of overwhelming evidence of Stephenson’s guilt, adequately demonstrates that Stephenson was prejudiced by his attorney’s failure to object to the stun belt. The panel’s conclusion that Stephenson’s claim fails for want of more concrete proof that he was harmed by the unobjected-to stun belt is inconsistent with the Supreme Court’s recognition that such evidence is impossible to marshal and unnecessary given the inherently prejudicial nature of visible restraints.
The court has remanded this case to the district court for further proceedings, and there remains the possibility that the dis*741trict judge might grant Stephenson relief as to the penalty phase of the trial, during which he also wore the stun belt. But whatever relief Stephenson might obtain as to the penalty phase will not address the prejudice he experienced vis-a-vis the jury’s assessment of his guilt. The proper course would be for this court to affirm the district court’s decision.
I respectfully dissent.
Appendix — Screen Shots from Petitioner’s Post Conviction Hearing Exhibit 45 News Video Depicting Stephenson
[[Image here]]
. Indeed, the Indiana Supreme Court found the use of stun belts so troubling that it prospectively banned their further use in Indiana *734courtrooms in Wrinkles, 749 N.E.2d at 1194-95.
. See Petitioner’s Postconviction Ex. 46 ¶ 3 ("During the trial, I became aware that John Stephenson was wearing a stun belt. I could see he had what appeared to be a rectangle shaped box attached to his lower back, underneath his shirt. I had seen a television show sometime previously [sic] to being selected for this jury that described what a stun belt was and what its purpose was. I have been hit by 220 volts and know what effect that amount of power has on a person. I believed the stun belt was to control John Stephenson’s behavior.”); Ex. 47 ¶ 3 ("I was aware that John Stephenson was wearing a device that would prevent him from running out of the courtroom. During the trial one of the male jurors mentioned to me that Mr. Stephenson was wearing this device. I was aware that he was wearing something that controlled his behavior.”); Ex. 48 ¶ 2 ("During the trial, I recall that Mr. Stephenson had some type of restraining device on him. I do not recall what the device looked like, or when or where I realized he [was] wearing one.”); Ex. 58 at 14 ("Q Did you know that John Stephenson had worn a stun belt through his trial? A I would say, yes, that I did because he wasn’t handcuffed, and naturally I assumed that he had that on because there was like a, you know, he wore like a loose shirt, basically like what you have on, loose, and there was a bulge back there in the back, so that's how I knew that.”).