dissenting:
Stevie Lamar Fields was charged with capital murder. He had the constitutional right to have twelve impartial jurors decide under California law whether he had committed the grave crime with which he was charged and, if so, whether he should lose his life as a result. Instead, he was sentenced to death by a jury whose foreperson brought into the jury room, and placed before his colleagues for consideration, lengthy Biblical quotations that clashed with the judge’s instructions, with California death penalty law, and with constitutional precepts .governing sentencing in a death penalty case. And he was convicted and sentenced to death by a jury containing one juror whose personal circumstances, objectively speaking and assuming an entirely good faith effort on his part to disregard those circumstances, made it highly unlikely that he could be the “impartial and indifferent” decision-maker that “due process alone has long demanded.” Morgan v. Illinois, 504 U.S. 719, 727, 112 S.Ct. 2222, 119 L.Ed.2d 492 (1992).
The federal district court held that Fields could not be condemned to death by a jury encouraged to rely on religious texts rather than the judge’s instructions as the basis for its fateful decision. The majority of this en banc court, however — after an exegesis that almost, but not quite, sanctions the jury’s collective recourse to lengthy quotations from the Bible — somehow divines that Fields would have been convicted and sentenced to death even if the jury had not consulted the Bible’s absolutist standards for imposing the death penalty. The majority also concludes that we must accept as dispositive a juror’s assertions that he was able to disregard the close resemblance between the unsolved kidnapping and rape of his wife not long before the trial and the crime with which Fields was charged. It does so even though the resemblance was so marked that the juror’s wife became convinced that Fields could have been her assailant and tried repeatedly during the trial to convince her husband to let her attend the trial so she could tell whether he was.1
Following the order in which the district court addressed Fields’s claims, I first examine the jury’s use of extrinsic materials — principally, passages from the Bible— during penalty phase deliberations. I explain why the district court correctly determined that this action violated Fields’s constitutional rights, and why the majority’s lack-of-prejudice analysis disregards the unusual problems that inhere in making a prejudice determination with regard to the impact of external influences on jury deliberations. I then address why, in the unique circumstances of this case — which go far beyond the fact that a juror’s wife was a rape victim and that Fields stood accused of rape — the challenged juror’s “potential for substantial emotional involvement, adversely affecting impartiality” is palpable. Tinsley v. Borg, 895 F.2d 520, 527 (9th Cir.1990) (quoting United States v. Allsup, 566 F.2d 68, 71 (9th Cir.1977)) (internal quotation marks omitted). This is therefore “one of Those extreme situations where the relationship between a ... juror and some aspect of the litigation is such that it is highly unlikely that the average person could remain impartial in his deliberations under the circumstances.’ ” Id. (quoting Person v. Miller, 854 F.2d 656, 664 (4th Cir.1988)).
*790I.
Before proceeding to either discussion, however, I begin with some reflections about the complex vision of the role of the jury in our legal system that underlies both sets of legal principles here applicable — those dealing with the sealing off of jurors from external influences once they are seated, and those dealing with juror bias or predisposition. Quite evidently, both sets of principles are grounded to some degree in a “black box” theory of the ideal jury: The perfect juror, on this posit, is a person who comes into court with all his or her reasoning processes intact, and no bias, predisposition, or prejudgment that would be an obstacle to fair decision-making. The perfect juror is then exposed only to that testimony, evidence, and argument the parties present and that the judge deems proper under the law, and is later told by the judge, and the judge alone, which legal principles to apply. Our paragon juror then limits his or her decision to that which was presented, that which was argued, and that which was instructed, generating a verdict pristinely insulated from all extraneous influences, internal and external.
But the “black box” theory of jury virtue is, quite obviously, far from the whole story. As the stress we place on obtaining a jury that represents a fair cross-section of a defendant’s peers indicates, see, e.g., Taylor v. Louisiana, 419 U.S. 522, 530-31, 95 S.Ct. 692, 42 L.Ed.2d 690 (1975), we also recognize, and value, the diverse perspectives jurors of different professions, racial backgrounds, economic circumstances, residential areas, and political and religious views bring to the determination of guilt and innocence and even, as in this case, life and death. Jurors all cut from the same mold, we have come to believe, are less likely to engage in useful collective deliberations, in which the whole is greater than its parts. Put another way, if all unbiased jurors of average intelligence were likely to think the same way with the same input, if jury deliberations were like arithmetic or algebra, then we could do with one juror rather than many. But we don’t, and we don’t want to. See Ballew v. Georgia, 435 U.S. 223, 230-39, 98 S.Ct. 1029, 55 L.Ed.2d 234 (1978) (holding that the Constitution requires criminal juries to comprise at least six members in part because a smaller size “leads to inaccurate fact-finding and incorrect application of the common sense of the community to the facts,” and because “the opportunity for meaningful and appropriate representation [of minority groups] does decrease with the size of the panels”); see also Tinsley, 895 F.2d at 528 (noting the foolishness of a doctrine that would categorically disqualify jurors from serving in cases in which their group affiliations might suggest a tendency toward a particular outcome (citing United States v. Salamone, 800 F.2d 1216, 1225 (3d Cir.1986))).
Other considerations as well temper in our jurisprudence the “black box” approach to jury deliberations. Among the most important is the emphasis we place on the importance of the privacy of jury deliberations. Jurors who expect that their deliberative processes will be open for exposure and interrogation after the verdict is in are likely to pull punches — to say less than they mean, to keep then-reasons to themselves and only pronounce bottom-line conclusions, and to fail to respond to points made by fellow jurors. See McDonald v. Pless, 238 U.S. 264, 267-68, 35 S.Ct. 783, 59 L.Ed. 1300 (1915) (“But let it once be established that verdicts solemnly made and publicly returned into court can be attacked and set aside on the testimony of those who took part in their publication and all verdicts could be, and many would be, followed by an inquiry in the hope of discovering something which might invalidate the finding.... If evi*791dence thus secured could be thus used, the result would be to make what was intended to be a private deliberation, the constant subject of public investigation; to the destruction of all frankness and freedom of discussion and conference.”).
An additional consideration is, of course, the recognition that jurors are simply human beings and human beings are not perfect, whether in their recollection, their understanding of language, or their ability fully to understand their own motivations and reasoning processes. See McDonough Power Equip., Inc. v. Greenwood, 464 U.S. 548, 555, 104 S.Ct. 845, 78 L.Ed.2d 668 (1984) (“The varied responses to respondents’ question on voir dire testify to the fact that jurors are not necessarily experts in English usage. Called as they are from all walks of life, many may be uncertain as to the meaning of terms which are relatively easily understood by lawyers and judges.”). Legal processes do have to come to an end, and litigants, while assuredly entitled to a fair trial, are not entitled to a perfect one. See id. (“To invalidate the result of a three-week trial because of a juror’s mistaken, though honest response to a question, is to insist on something closer to perfection than our judicial system can be expected to give.”).
This constellation of ideals, competing values, and practical considerations has led to a nuanced set of procedures and standards designed to assure an “impartial and indifferent” jury without losing sight of the value of diversity of backgrounds, the need for insulating the jury’s deliberative process in large degree from later inquiry, and the reality that human perfection is not likely any time in the near future. Those procedures include voir dire, designed to ferret out without unduly intrusive inquiry those individuals who have a connection to the particular ease — to its parties, its facts, its legal standards — such that, consciously or not, that connection is simply too likely to be a barrier to a fair— not pristine but fair — consideration of the evidence, argument, and legal standards presented. They also include instructions to the jury forbidding them to discuss the case with anyone other than fellow jurors once seated and not to discuss the case with even fellow jurors before the instructions are given and the jury retires to deliberate. See Cal. JüRY InstR. Crim. 0.50 (“You must not converse among yourselves, or with anyone else, including but not limited to, spouses, spiritual leaders or advisers, or therapists, on any subject connected with the trial, except when all the following conditions exist: (a) The case has been submitted to you for your decision by the court, following arguments by counsel and jury instructions; (b) You are discussing the case with a fellow juror; and (c) All twelve jurors [and no other persons] are present in the jury deliberating room.” (brackets in original)). At the same time, only rarely do we insist that jurors must be cloistered altogether during trial, rather than going home to their communities— and to the real possibility of exposure to media, to the opinions of friends and family, and to the opportunity to conduct investigations or legal inquiries beyond those available in the courtroom. And the accommodating jury procedures also include evidentiary limitations on inquiry into the jury’s verdict once it issues — limitations that are not absolute, but which are designed to allow inquiry only into objective, external factors that may have interfered with the jury’s functioning. See Fed. R. Evid. 606(b) (“Upon an inquiry into the validity of a verdict ..., a juror may not testify as to any matter or statement occurring during the course of the jury’s deliberations or to the effect of anything upon that or any other juror’s mind or emotions as influencing the juror to assent to or dissent from the verdict ..., or concerning the juror’s mental processes in connection therewith. But a juror may *792testify about (1) whether extraneous prejudicial information was improperly brought to the jury’s attention, [and] (2) whether any outside influence was improperly brought to bear upon any juror.... ”).
All three of these sets of procedures could be much more absolute if our commitment to the “black box” jury were rigid — which, as I have said, it decidedly is not. Yet, they all posit a limit beyond which an extreme departure from the ideal is unacceptable and will lead to reversal of a verdict. That limit is reached, in the most general of terms, when the circumstances of either one or more jurors or of the nature of the jury deliberations are such that we simply lose confidence that the verdict was reached on the basis of the facts, argument, and legal standards presented in the courtroom.
The majority places a great deal of its emphasis upon the considerable barriers we have erected, for very good reason as I have said, upon post-hoc inquiry into jury verdicts, and upon our encouragement of interchange among jurors based on their life experiences. In doing so, however, it loses sight of the fact that we have not let go of the conviction that there are circumstances in which the connection of a juror to the particulars of a case is so great that an emotional rather than rational verdict is likely — and is likely whether the juror so recognizes or not — because, as the very juror in question in this case said, “you can never be sure what’s in the back of your mind.” And it also loses sight of the impermeable line we have set between drawing on one’s life experience and active research of outside sources, with regard to the facts or to the legal standards that are to govern the case.
The distinctions drawn may appear fine, but they are established and they are the result of considered compromises reached over time between the competing considerations that govern jury deliberations, some of which I have suggested. To resolve this case on the basis of slogans — for example, the assertion that jurors bring their moral precepts to the jury room, or that we must believe jurors who say they can overcome any emotional relationship they have to the particular facts — is simply to disregard the careful balances struck in our case law, so to allow the all-important jury system to serve its critical purposes.
II.
A.
Before beginning penalty-phase deliberations, the jury was instructed that “[a]fter having considered all of the evidence in this case and having taken into account all of the applicable factors upon which you have been instructed,[2] you shall deter*793mine whether the penalty to be imposed on defendant shall be death or confinement in the State Prison for life without the possibility of parole.” Nonetheless, after the first day of penalty-phase deliberations in Fields’s trial, the jury foreperson, Rodney White, went home; consulted the Bible and a dictionary; wrote out three pages of notes — including verbatim copies of three Bible passages, Genesis 9:6, Exodus 21:12, and Romans 13:1-5, widely understood to advocate capital punishment; brought those notes into the jury room the next morning; and shared them with his fellow jurors.
We have consistently recognized that the Sixth Amendment prohibits jurors from introducing matters into deliberations not presented during the trial. See Gibson v. Clanon, 633 F.2d 851, 854 (9th Cir.1980) (explaining that a jury’s consideration of extrinsic material is a constitutional violation). Although our case law often refers to such misconduct as the jury’s consultation of “extrinsic evidence” or “extraneous facts,” we have explained that “[ejxtraneous-evidence cases involve not only the introduction of ‘evidence’ per se but the ‘submission of “extraneous information” (e.g., a file or dictionary) to the jury.’ ” United States v. Rosenthal, 454 F.3d 943, 949 (9th Cir.2006) (quoting United States v. Madrid, 842 F.2d 1090, 1093 (9th Cir.1988)); see also Marino v. Vasquez, 812 F.2d 499, 502-03, 505 (9th Cir.1987) (holding that consulting a dictionary definition for the meaning of “malice” constituted the consideration of extrinsic information). Instead, analysis of an extrinsic information claim depends on whether the outside information “pertainfs] to ‘any fact in controversy or any law applicable to the case.’ ” Madrid, 842 F.2d at 1093 (emphasis added) (quoting Rushen v. Spain, 464 U.S. 114, 121, 104 S.Ct. 453, 78 L.Ed.2d 267 (1983) (per curiam)); see also Thompson v. Borg, 74 F.3d 1571, 1574 (9th Cir.1996) (“Juror misconduct typically occurs when a member of the jury has introduced into its deliberations matter which was not in evidence or in the instructions.” (emphasis added)).
Here, there is no question that the Biblical passages copied by White pertained to the key legal question before the jury at the penalty phase — -whether death was the appropriate sentence for Fields’s acts. As the district court correctly observed, several of the passages expressed an absolute command to execute murderers:
The Biblical passages cited by Juror White were not general passages dealing with morality or the commonplace principle that capital punishment is permissible in the abstract in the Judeo-Chris-tian ethical and religious tradition. The references directed the jury that the death penalty should be imposed in any case involving murder.
Fields v. Calderon, No. CV 92-0465 DT, slip op. at 15 (C.D.Cal. Jan. 18, 2000) (emphasis added) (citation and internal quotation marks omitted).
There is also no question that in their absolute nature, the Biblical passages contained in White’s notes clashed with standards California law provided for making such decisions. The jurors were instructed to make their decision based on California’s 1977 death penalty statute that detailed an individualized process for de*794termining whether a defendant should be executed:
After having heard and received all of the evidence, the trier of fact shall consider, take into account and be guided by the aggravating and mitigating circumstances referred to in this section,[3] and shall determine whether the penalty shall be death or life imprisonment without the possibility of parole.
Cal. Penal Code § 190.3 (1977). The California Supreme Court has held this statute “require[d] the jury to concentrate upon the circumstances surrounding both the offense and the offender.” People v. Jackson, 28 Cal.3d 264, 316, 168 Cal.Rptr. 603, 618 P.2d 149 (1980), overruled on other grounds by People v. Cromer, 24 Cal.4th 889, 103 Cal.Rptr.2d 23, 15 P.3d 243 (2001). Accordingly, “religious doctrine, commandments or biblical passages” are “factors outside section 190.3.” People v. Sandoval, 4 Cal.4th 155, 193-94, 14 Cal.Rptr.2d 342, 841 P.2d 862 (1992), aff'd on other grounds, 511 U.S. 1, 114 S.Ct. 1239, 127 L.Ed.2d 583 (1994).
By introducing the absolutist Biblical commands into deliberations, White effectively suggested that the jury ignore the individualized sentencing process provided by state law and demanded by the federal
Constitution. As we said in Sandoval v. Calderon:
[I]nvocation of higher law or extra-judicial authority violates the Eighth Amendment principle that the death penalty may be constitutionally imposed only when the jury makes findings under a sentencing scheme that carefully focuses the jury on the specific factors it is to consider in reaching a verdict. The Biblical concepts of vengeance ... do not recognize such a refined approach. Argument involving religious authority also undercuts the jury’s own sense of responsibility for imposing the death penalty.
241 F.3d 765, 776-77 (9th Cir.2001) (citations omitted);4 see also Robinson v. Polk, 444 F.3d 225, 232 (4th Cir.2006) (King, J., dissenting from the denial of rehearing en banc) (“In effect, this juror requested that his fellow jurors throw the individualized consideration required by the Constitution to the wind, for while the Constitution requires that the death penalty be imposed through structured discretion on only a narrow class of the worst murderers, the principle of ‘an eye for an eye’ licenses death as a punishment for any murder, a position rejected by the Supreme Court as contrary to the Constitution.”).5 Substan*795tively, then, there is no doubt that the Biblical quotations introduced a set of standards that contradicted the ones the jury was supposed to apply in deciding whether Fields was to live or die.
As to the procedural propriety of consulting the Bible during deliberations, federal and state appellate courts generally agree when engaging in de novo review, that a jury engages in the unconstitutional consultation of extrinsic material by introducing the Bible into deliberations during a capital trial. McNair v. Campbell, 416 F.3d 1291, 1308 (11th Cir.2005) (recognizing “it is undisputed that jurors ... considered extrinsic evidence during their deliberations” when the jury foreperson read aloud from a Bible), cert. denied, 547 U.S. 1073, 126 S.Ct. 1828, 164 L.Ed.2d 522 (2006);6 Jones v. Kemp, 706 F.Supp. 1534, 1559 (N.D.Ga.1989) (holding the use of a Bible by the jury constituted an impermissible “search for the command of extrajudicial ‘law’ from [a] source other than the trial judge”); McNair v. State, 706 So.2d 828, 837 (Ala.Crim.App.1997) (analyzing a jury’s use of the Bible during deliberations according to the “well settled principle of law ... [that] is fundamental to a fair trial ... that jurors should consider only the evidence presented at trial” (quoting Ex parte Troha, 462 So.2d 953, 954 (Ala.1984)) (internal quotation mark omitted)); People v. Banks, 32 Cal.4th 269, 308, 8 Cal.Rptr.3d 767, 82 P.3d 1249 (2004) (holding a juror engaged in misconduct by “bringing a copy of the [Biblical] passage into the jury room, and passing it around to the other jurors”); People v. Harlan, 109 P.3d 616, 629 (Colo.2005) (holding the use of written Biblical materials in the jury room was improper under state law that prohibited “[e]xposure of a jury to information or influences outside of the trial process itself’); State v. Harrington, 627 S.W.2d 345, 350 (Tenn.1981) (holding error occurred when “the jury foreman buttressed his argument for imposition of the death penalty by reading to the jury selected biblical passages”); Lenz v. Warden of the Sussex I State Prison, 267 Va. 318, 593 S.E.2d 292, 298-99 (2004) (analyzing defendant’s claim that a Bible was present in the jury room pursuant to the Supreme Court’s test for improper extraneous jury contacts).7 Although several of these cases involve the presence of a complete copy of the Bible in the jury room, making only certain portions available exacerbates, rather than ameliorates, the problem presented by the introduction of Biblical writings during jury deliberations. Here, for example, the selection leaves out Biblical passages that can be read as condemning the death penalty or as condoning mercy in some instances. See Dissent of Gould, J. at 785-86 n.2.
Unable to rely on case law, the majority suggests that White’s conduct was not misconduct because the notes “are notions of general currency that inform the moral judgment that capital-case jurors are called upon to make.” Maj. op. at 780. *796Initially, I note that the majority’s observation would likely be challenged by tens of millions of Americans who view the Bible not as a collection of “notions” about moral principles, but as a repository of hard-and-fast imperatives that must direct daily life. See Robinson v. Polk, 438 F.3d 350, 374 (4th Cir.) (King, J., dissenting) (“[T]he majority ignores the fact that the Bible is an authoritative code of morality— and even law — to a sizable segment of our population.”), cert. denied — U.S. -, 127 S.Ct. 514, 166 L.Ed.2d 383 (2006). White’s notes were therefore significant for their factual representation that the Bible contained such statements, apart from the moral philosophy that the statements themselves expressed.
In any ease, I fail to understand why a distinction between extrinsic statements of general moral currency and other extrinsic materials has a legal bearing on this case. Although the majority suggests that the distinction is significant in this case because the jury in a capital trial is assigned the task of making a moral judgment, as I have explained, the absolute nature of Biblical materials introduced by White contravened the permitted role of moral considerations and instead violated the principle that capital sentencing must be individualized. Moreover, the moral nature of death penalty judgments does not allow religious considerations to be a proper matter for deliberations, rather than a factor jurors can privately contemplate in the course of undertaking their awesome responsibility. See Robinson, 444 F.3d at 227 (Wilkinson, J., concurring in the denial of rehearing en banc) (“There is a difference between a juror bringing a Bible into the jury room for personal strength and support and the jury as a whole reading and debating the biblical text as the basis for a life and death decision. Such a debate is constitutionally problematic.... If the presence of a Bible in the jury room drives the collective discussion, and renders a capital sentence the result of religious command, then in my view, an important line has been crossed.”).
The majority also suggests that all the courts that have held consulting the Bible to be impermissible reliance on extrinsic material are wrong because the Biblical quotations White looked up, copied, and brought into the jury room were simply “general, commonly known points in favor of the death penalty,” and “[i]t is difficult to see how sharing notes can be constitutionally infirm if sharing memory isn’t.” Maj. op. at 780. But this suggested equivalence disregards the careful balance between the various precepts regarding jury deliberations I discussed at the outset. In fact, as Morgan makes clear, a juror who voted for the death penalty on the basis of the absolutist position sanctioned by the Biblical quotations White placed before his fellow jurors would violate his oath to follow California law. For quite separate reasons — principally, preservation of the privacy of jury deliberations and of the finality of jury verdicts — we would not allow inquiry into why any individual juror voted for the death penalty, or into statements made to others regarding why they should do the same. But that forbearance serves an independent interest; it does not sanction the disregard of the instructions.
There is a second reason why this case is not similar to ones involving a juror introducing an argument into deliberations based on his personal knowledge. In such cases, we have held that no misconduct occurred. See Rodriguez v. Marshall, 125 F.3d 739, 745 (9th Cir.1997) (juror discussed difficulty he had “discerning and recalling objects while driving at freeway speeds”); McDowell v. Calderon, 107 F.3d 1351, 1367 (9th Cir.) (juror argued during deliberations that “a sentence of life without parole ... wouldn’t mean ‘without pa*797role’ ”), vacated en banc in other parts, 130 F.3d 833 (9th Cir.1997); Hard v. Burlington N. R.R. Co., 870 F.2d 1454, 1462 (9th Cir.1989) (juror made arguments during deliberations based on his prior military experience interpreting x-rays). But these cases are based on the proposition that “the general knowledge, opinions, feelings, and bias that every juror carries into the jury room ” are properly considered during deliberations. Hard, 870 F.2d at 1461 (emphasis added); see also United States v. Navarro-Garcia, 926 F.2d 818, 821 (9th Cir.1991) (“[A] juror’s past personal experiences may be an appropriate part of the jury’s deliberations.” (emphasis added)). As I have noted, the success of the jury system rests in large part on the coming together of these variations in background, perception, and point of view.
In this case, however, it is undisputed that White’s notes were the product of overnight Biblical research, rather than of familiarity with the Bible that White already had at the outset of deliberations. The case law proscribing importation of external information places its all on the proposition that this distinction matters: After we choose jurors, we want the decision made on the basis of what went on in the courtroom, filtered through the personalities, background information, and reasoning ability the jurors brought with them to court. But we do not approve of, and regard as misconduct, affirmatively gathering outside information. Further, the lengthy quotations, written down and passed around, conveyed a sense of authority quite different from a paraphrase or one line quotations spoken from memory, not least because they could be consulted repeatedly and outside of White’s immediate presence. The written, lengthy quotations introduced tangibly an external exhortation, that God, or the authors of God’s book — not just juror White — encouraged jurors to disregard the judge’s instructions and vote for the death penalty no matter what. The majority, therefore, is wrong to portray this case as similar to “sharing memory,” and this case does not raise the question of whether jurors act impermissibly by referring to their faith during deliberations. See Maj. op. at 780.
B.
Ultimately, however, the majority equivocates about whether White engaged in misconduct, but holds that his actions did not prejudice Fields. In so doing, the majority ignores the indications deemed relevant by our established case law, and fails to appreciate the unique issues involved in examining an extrinsic information claim in the context of an attack on a jury verdict.
1. The district court determined that White’s misconduct did prejudice Fields. Before so concluding, the court made critical factual findings concerning the introduction of the Biblical material into the jury room: The district court found that the introduction came at a time when the jury was divided on the proper sentence.
[T]he jurors considered and discussed the Biblical references in their deliberations. Juror Henry stated that, the foreman brought to the deliberations pages of notes of citations from the Bible and other religious sources which he felt supported capital punishment. These notes were passed around to and discussed by the jury. It was only after we reviewed and discussed the notes that an unanimous decision in favor of death was reached.
Juror Hilliard stated that, “[t]he jury foreman presented to us, and we discussed, information which he had brought from home, including excerpts from the Bible and definitions. It was after these discussions that we were able to reach a unanimous verdict in *798favor of recommending the imposition of the death penalty.” Juror White stated that he “brought the notes to the penalty phase jury deliberations and the contents of these notes were discussed during our deliberations.”
Respondent has submitted declarations from several jurors in which they stated that they did not recall any discussion of the Bible or dictionary definitions. However, the declarations submitted by respondent confirm the Court’s finding that Biblical references were provided by Juror White and discussed by the jury.
... [I]n this case, there is evidence that a majority of the jurors favored a verdict of life without the possibility of parole until the jury discussed the Biblical references.
Fields v. Calderon, slip op. at 13-14, 16-17 (second alteration in original) (citations omitted).
The district court quite properly considered the jurors’ statements that the jury was undecided at the time the Biblical material was brought into the jury room and that unanimity came only after that point. Under Rule 606(b) of the Federal Rules of Evidence, federal courts can consider “juror testimony about the consideration of extrinsic evidence” but cannot consider testimony “about the subjective effect of evidence on the particular juror.” Sassounian v. Roe, 230 F.3d 1097, 1108 (9th Cir.2000). Here, the district court did not find that the discussion of the Biblical material was the reason that jurors changed their vote, but merely that vote changes occurred, and when. We have previously found it proper to consider the timing of shifts in jury votes relative to the introduction of extrinsic evidence. See id. at 1110; Marino, 812 F.2d at 505 & n. 8; see also Mattox v. United States, 146 U.S. 140, 147-51, 13 S.Ct. 50, 36 L.Ed. 917 (1892) (holding, after examining the common law restrictions against impeaching verdicts through juror testimony, that the defendant should receive a new trial based on jury misconduct because admissible evidence demonstrated that “[t]he jury in the case before us retired to consider of their verdict on the 7th of October, and had not agreed on the morning of the 8th, when the newspaper article was read to them”).8
Moreover, that White thought it necessary to bring the Biblical material into the jury room after an initial session of deliberations strongly suggests that the material could have had an impact on the jury. Presumably, White felt that reviewing the material at home had affected his analysis and thought it might have a similar impact on others if brought into the jury room. Why would he spend the time hand copying the material if the jury was close to a decision, in which case his work was likely to be for naught? See Gibson, 633 F.2d at 855 (“[T]he fact that at least two jurors believed that it was necessary to obtain more evidence is, by itself, an indication that there may have been a need to resolve some lingering hesitation or uncertainty.”).
Also, quite aside from the time and manner in which this extrinsic information was introduced, its content had a clear potential to affect deliberations. We have previously observed that one of the Bible passage reprinted verbatim in White’s notes — Romans 13:1-5 — is “commonly understood as providing justification for the imposition of the death penalty,” and its invocation during the sentencing phase of a capital trial “cloak[s] the State with God’s authori*799ty.” Sandoval, 241 F.3d at 775, 779. We held in Sandoval that a prosecutor’s allusion to that passage, in concert with other religious references, prejudiced a defendant’s right to be sentenced according to the statutory scheme for imposing the death penalty. Id. at 778-80. Sandoval explicitly noted that the record did not disclose whether the jury actually considered the prosecutor’s Biblical argument, but it nonetheless held that “we cannot assume that the prosecutor’s religious argument did not persuade at least one of the jurors to change a vote for life to death.” Id. at 779.
Notwithstanding the majority’s attempt to frame a prosecutor’s invocation of the Bible as more damaging than its actual entry into the jury room, Maj. op. at 780-81, our case law and common sense dictate the opposite conclusion: A fellow juror’s introduction of such material into the jury room has an even greater potential for a prejudicial effect, because the defendant is unable to mitigate the jury’s consideration of the Bible as he can when the prosecutor brings it into the trial — for example, by tailoring his closing argument to account for the religious arguments or by insuring that the judge instructs the jury to consider only the relevant statutory factors. See Gibson, 633 F.2d at 854 (“[W]hen a jury considers facts that have not been introduced in evidence ... the violation may be more serious than where these rights are denied at some other stage of the proceedings because the defendant may have no idea what new evidence has been considered. It is impossible to offer evidence to rebut it, to offer a curative instruction, to discuss its significance in argument to the jury, or to take other tactical steps that might ameliorate its impact.”). Moreover, we do not know whether any juror in Sandoval paid any attention to the prosecutor’s Biblical references. But we do know that at least one juror in this case— White — was sufficiently concerned about what the Bible said about the death penalty to spend time copying out lengthy quotations and that other jurors reviewed the passages. So we have here direct evidence of an impact on the jury deliberations that was lacking in Sandoval.
Other courts have also recognized the specially prejudicial nature of a jury’s consideration of Biblical material. One federal district court has held that a jury’s consultation of a Bible during penalty phase deliberations “may be highly prejudicial to the defendant” because it represents “a source which ‘would likely carry weight with laymen and influence their decision.’ ” Jones, 706 F.Supp. at 1560 (quoting Wilson v. Kemp, 777 F.2d 621, 626 (11th Cir.1985)). Likewise, Colorado’s Supreme Court has found prejudice from the introduction of the Bible into deliberations: “[A]t least one juror in this case could have been influenced by these authoritative passages to vote for the death penalty when he or she may otherwise have voted for a life sentence” because “[t]he Bible and other religious documents are considered codes of law by many in the contemporary communities from which ... jurors are drawn.” Harlan, 109 P.3d at 630-31.
2. Given the nature of the extrinsic information involved in this case and the manner in which it was received by the jury, the majority’s, reversal of the district court on no-prejudice grounds is inexplicable. In so ruling, the majority determines that the notes had “no substantial and injurious effect or influence in determining the jury’s verdict” and holds this lack of impact precludes habeas relief pursuant to Brecht v. Abrahamson, 507 U.S. 619, 113 S.Ct. 1710, 123 L.Ed.2d 353 (1993). Maj. op. at 781.
The Supreme Court’s requirement in Brecht that habeas relief be granted only when an error has a “substantial and inju-*800nous effect” was derived from the harmless-error standard enunciated in Kotteakos v. United States, 328 U.S. 750, 66 S.Ct. 1239, 90 L.Ed. 1557 (1946). 507 U.S. at 623, 113 S.Ct. 1710. Kotteakos explains our task in conducting this variety of harmless-error review:
If, when all is said and done, the [court’s] conviction is sure that the error did not influence the jury, or had but very slight effect, the verdict and the judgment should stand.... But if one cannot say, with fair assurance, after pondering all that happened without stripping the erroneous action from the whole, that the judgment was not substantially swayed by the error, it is impossible to conclude that substantial rights were not affected. The inquiry cannot be merely whether there was enough to support the result, apart from the phase affected by the error. It is rather, even so, whether the error itself had substantial influence. If so, or if one is left in grave doubt, the conviction cannot stand.
328 U.S. at 764-65, 66 S.Ct. 1239; see also Payton v. Woodford, 346 F.3d 1204, 1218 (9th Cir.2003) (en banc) (explaining that our Brecht inquiry must reflect the “greater need for reliability” in the death penalty context (quoting Coleman v. Calderon, 210 F.3d 1047, 1050 (9th Cir.2000)) (internal quotation marks omitted)), rev’d on other grounds sub nom. Brown v. Payton, 544 U.S. 133, 125 S.Ct. 1432, 161 L.Ed.2d 334 (2005). Where the contention is that the jury improperly considered extrinsic information, this Brecht inquiry must focus upon the extrinsic material’s impact on any juror, because “even a single juror’s improperly influenced vote deprives the defendant of an unprejudiced, unanimous verdict.” Lawson v. Borg, 60 F.3d 608, 613 (9th Cir.1995); see also Sassounian, 230 F.3d at 1110.
This circuit for more than two decades has relied on a five-factor test to determine in the habeas context whether extrinsic information is prejudicial. That test requires us to consider:
(1) whether the material was actually received, and if so, how; (2) the length of time it was available to the jury; (3) the extent to which the jury discussed and considered it; (4) whether the extrinsic material was introduced before a verdict was reached, and if so, at what point in the deliberations it was introduced; and (5) any other matters which may bear on the issue....
Bayramoglu v. Estelle, 806 F.2d 880, 887 (9th Cir.1986) (citing Paz v. United States, 462 F.2d 740, 746 (5th Cir.1972)). I, like the district court, would hold that all of these factors point toward the prejudicial nature of White’s notes.
First, there is no dispute that the material was actually received by the jury when its foreperson brought the notes into deliberations. Second, the notes entered the jury room at the start of the second day of deliberations, which meant the jury considered the Biblical material during the majority of the time it considered Fields’s sentence — four-and-one-half of the six-and-one-half hours of deliberations. Third, the notes were available to all members of the jury, and several members of the jury submitted affidavits attesting to their presence in the jury room. Fourth, the material entered the jury room during the pivotal period when the sentence was far from settled. Fifth, “the extrinsic information directly related to a material issue in the case” — the proper sentence. Lawson, 60 F.3d at 612; see also Marino, 812 F.2d at 506 (“[R]eversible error commonly occurs where there is a direct and rational connection between the extrinsic material and a prejudicial jury conclusion, and where the misconduct relates directly to a material aspect of the case.”).
*801Moreover, only one of the factors we identified in Jeffries v. Wood, 114 F.3d 1484 (9th Cir.1997) (en banc), as potentially diminishing the prejudice created by the jury’s consideration of extrinsic information is even arguably present in this case. Those factors are
whether the prejudicial statement was ambiguously phrased; whether the extraneous information was otherwise admissible or merely cumulative of other evidence adduced at trial; whether a curative instruction was given or some other step taken to ameliorate the prejudice; the trial context; and whether the statement was insufficiently prejudicial given the issues and evidence in the case.
Id. at 1491-92 (footnotes omitted).
The majority, in determining that the extrinsic material here does not meet the Brecht prejudice standard, judges the prejudicial nature of White’s misconduct based on a set of factors it creates largely from thin air: the fact that White’s notes also contained material helpful to Fields; the fact that the extrinsic material was introduced early in the deliberations; the fact that the jury was instructed to base its determinations on the law; and the fact that the aggravating evidence against Fields was powerful. Maj. op. at 780-81. These reasons, with one exception, have no basis in our case law and are, moreover, not pertinent to the prejudice inquiry.
Far from providing “fair assurance” about the harmlessness of the jury misconduct, the majority’s first three factors provide no reason to doubt that the Biblical material affected deliberations. The majority’s first factor, the balance of White’s notes, ignores the record: As the district court found, “all of the Biblical references supported the imposition of the ultimate penalty” and “did not include ... Biblical references supporting the concepts of forgiveness and mercy.” Fields v. Calderon, slip op. at 16. So the effect of the Biblical material was exclusively in favor of capital punishment. White’s “against” notes were much shorter, were not quotations at all— let alone lengthy Biblical quotations-and reflected no external research.
The majority’s second factor, the early introduction of the Biblical materials into deliberations, actually made the misconduct worse, because the jury engaged in very little deliberation unaffected by the religious considerations. See Lawson, 60 F.3d at 613 (finding “the early stage at which the extrinsic information was introduced” was one factor dictating that “the juror misconduct substantially and injuriously influenced the verdict”). The majority’s third factor, our normal presumption that jurors follow instructions, is inappo-site when the jury’s very decision to consider the extrinsic evidence conclusively demonstrates that it did not follow the instruction to rely only on the law as stated by the judge. Also, this factor would render all jury misconduct concerning extrinsic material harmless, because jurors are uniformly instructed to rely on the facts and on legally proper considerations.
Ultimately, the only potentially relevant factor cited by the majority for its prejudice analysis is the presence of powerful aggravating evidence in this case. Certainly, the murder charged in this case was extremely brutal and was alleged to have occurred during the course of an extensive violent crime spree. Based on such evidence, I have no doubt that a jury could have determined death was the proper punishment for Fields after properly applying California’s prescribed weighing system.
Under Brecht, however, “[t]he inquiry cannot be merely whether there was enough to support the result, apart from the phase affected by the error. It is *802rather, even so, whether the error itself had substantial influence.” Kotteakos, 328 U.S. at 765, 66 S.Ct. 1239. The manner in which the jury considered the extrinsic Biblical material in this case, as well as its inherently prejudicial nature, demonstrates that the error did substantially influence this jury and that Fields has a right to be resentenced by a jury not so influenced. The California Supreme Court’s analysis of how a hypothetical jury would have responded to alternate defense strategies, the basis for the prejudice analysis that the majority opinion quotes at length, is not instructive.
3. Not only is the majority’s Brecht analysis flawed as applied to this specific case, but it ignores the special nature of claims involving jury misconduct. On direct review of federal criminal cases, after a defendant has demonstrated that the jury considered extrinsic evidence, we readily find prejudice absent specific evidence that there was no prejudice. See Rosenthal, 454 F.3d at 949 (“Extraneous information cases ... call for more searching review; we grant a new trial if ‘there is a reasonable possibility that the material could have affected the verdict.’ ... [W]e generally place the burden ‘on the party opposing .a new trial to demonstrate the absence of prejudice.’ ” (emphasis omitted) (quoting Sea Hawk Seafoods, Inc. v. Alyeska Pipeline Serv. Co., 206 F.3d 900, 906 (9th Cir.2000))). This view of prejudice arises from the Supreme Court’s presumption of prejudice in Mattox and in Remmer v. United States, 347 U.S. 227, 74 S.Ct. 450, 98 L.Ed. 654 (1954), when jurors communicate with third parties. See United States v. Martinez, 14 F.3d 543, 550 (11th Cir.1994) (citing Remmer, 347 U.S. at 229, 74 S.Ct. 450; Mattox, 146 U.S. at 150, 13 S.Ct. 50).9
Even without applying a legal presumption, finding prejudice in cases where a defendant has proved that' the jury received extrinsic information somewhat more readily than one might in other instances is quite sensible, both on direct review and in habeas. Rule 606(b) of the Federal Rules of Evidence makes it uniquely difficult to prove that extrinsic evidence had an actual effect on jurors. That rule prohibits jurors from testifying “about the subjective effect of [extrinsic] evidence on the particular juror.” Sassounian, 230 F.3d at 1108. So a court will never have admissible evidence directly linking extrinsic materials with a juror’s final vote, see United States v. Rutherford, 371 F.3d 634, 644 (9th Cir.2004) (explaining that under Rule 606(b) “a juror cannot *803testify to whether an outside influence caused him to change his vote from innocent to guilty”). As the Eighth Circuit has observed in explaining the Mattox/Remmer presumption:
Because Rule 606(b) precludes the district court from investigating the subjective effects of any extrinsic material on the jurors, whether such effects might be shown to affirm or negate the conclusion of actual prejudice, a presumption of prejudice is created and the burden is on the government to prove harmlessness.
United States v. Bassler, 651 F.2d 600, 603 (8th Cir.1981).
The test we adopted in Bayramoglu for determining whether the jury’s consideration of extrinsic information is prejudicial has effectively served as our means of mediating between the right to secrecy and finality under Rule 606(b) and a habeas petitioner’s constitutional right to a jury untainted by extrinsic evidence. See Sassounian, 230 F.3d at 1109 (noting Rule 606(b) requires the court “to ignore the most direct evidence of prejudice ... lending] an ‘Alice in Wonderland quality to the discussion of whether[the defendant] was actually prejudiced by the admitted jury misconduct,’ ” but holding that the Bayramoglu, factors led to a determination . that the defendant’s rights were violated notwithstanding the Rule 606(b) limitation (quoting People v. Sassounian, 182 Cal.App.3d 361, 419, 226 Cal.Rptr. 880 (1986) (Johnson, J., dissenting))); United States v. Castello, 526 F.Supp. 847, 849-50 (W.D.Tex.1981) (“The rule is well established that a jury may not impeach its own verdict. The Court is not at liberty to investigate ‘the subjective effects of any breach on any jurors, whether such effects might be shown to affirm or negate the conclusion of actual prejudice.’ Instead, the Court must conduct an inquiry into the prejudicial potential of extraneous material on the average juror. Objective facts, therefore, become the focus of the inquiry.” (quoting Simon v. Kuhlman, 488 F.Supp. 59, 68 (S.D.N.Y.1979)) (citations omitted)), cited in Bayramoglu, 806 F.2d at 887. The majority’s decision not to apply this test in conducting its Brecht analysis upsets our settled balance for determining whether “even a single juror’s” vote was “improperly influenced” — the relevant constitutional question, Lawson, 60 F.3d at 613 — when courts are prohibited from directly finding out from jurors the effect of extrinsic information on their deliberative process.
* * *
In sum, I have no doubt that White engaged in unconstitutional misconduct by injecting his overnight Biblical research into the deliberations, and I am convinced, taking into account the manner in which the material came before the jury, its nature, and our usual test for determining prejudice arising from extrinsic material, that this conduct substantially influenced the jury’s penalty phase deliberations. Fields was deprived of his right to have his fate decided by jurors who applied only the applicable legal standards, rather than conflicting standards derived from scripture.10
*804III.
Fields was not only denied his right to an impartial jury applying the correct law, as the district court found. In addition, the presence of an individual on the jury who, objectively speaking, was not “impartial and indifferent,” violated Fields’s constitutional rights even before the jury considered the Biblical material during its penalty-phase deliberations. See Morgan, 504 U.S. at 727, 112 S.Ct. 2222.
The wife of juror Floyd Hilliard had been the victim of an unsolved crime, with details quite similar to the allegations against Fields. The extreme distress generated by the attack — stoked by the fear that the unknown perpetrator would return — was still affecting the family at the time of the trial. Moreover, during the trial Floyd Hilliard was faced with repeated suggestions from his wife that Fields was her rapist, and had to respond nightly to her requests to attend the trial to confirm her suspicions.
Any of these circumstances, independently, would likely place unusual and unpredictable strains on Floyd Hilliard’s objectivity. Their combination presents the “extraordinary case[], [in which] courts may presume bias based on the circumstances.” Dyer v. Calderon, 151 F.3d 970, 981 (9th Cir.1998) (en banc). Because the strain from the combination of these facts was far too likely to affect Floyd Hilliard’s deliberative process whether he so recognized or not, his presence on the jury violated Fields’s constitutional rights.
A.
Two-and-one-half years before Floyd Hilliard was called for jury duty, his wife Diane Hilliard, driving home from a Christmas party, had been abducted at gunpoint by a man who rammed her car while it was idling at a traffic light. The assailant forced Diane Hilliard into his car, drove her to a remote location, and pistol-whipped, robbed, and raped her. He then told her, “well, bitch, I’m going to have to shoot you now.” His plan, however, was interrupted by a local resident who noticed the suspicious car and came out to investigate. Diane Hilliard recounted that the attacker “threatened to come back to finish me off’ before fleeing — a quite plausible threat because he had stolen her purse that contained her home address. .When Floyd Hilliard arrived at the hospital after the attack, he found his wife disheveled and in a state of shock, nursing a broken nose, head lacerations, and a black eye. According to his wife, Floyd Hilliard was “shaken,” “upset,” “shocked,” and “angry” upon seeing her injured state. More than twenty-five years later, Diane Hilliard had no hesitation in calling the attack “the most horrifying experience in my life.”
According to Diane Hilliard, the crime “radically changed” both her and her husband’s life, in part because the perpetrator was never caught. In a 2003 deposition, she confirmed that “[f]or a long time after the attack[,] I lived in terror that he was out there” and feared “that he would come to make good on his threat.” Floyd Hilliard also confirmed in a 2003 deposition that the attack was “a very, very upsetting event.” Following the attack, Diane Hilli-ard began carrying a handgun in her purse, and continued to do so during the time of Fields’s trial. And, mirroring his wife’s fear, Floyd Hilliard for the first couple of weeks after the attack frequently sat by the window of his house with a shotgun, watching for the rapist’s return. Floyd Hilliard testified that if the attacked “had come back, I had something for him.... If he made unauthorized entry, enter at your own risk.” Diane Hilliard testified that her fear began to subside only when the family moved to a new home, several years after Fields’s trial.
The allegations against Fields included many of the same crimes — rape, kidnap*805ping, robbery, assault — as those suffered by Diane Hilliard at the hands of a perpetrator still on the loose. Moreover, the details and modus operandi of those crimes had substantial similarities. Diane Hilli-ard was attacked by a slender African American male in his twenties, which matched the physical description of Fields. She was abducted while driving through a neighborhood only ten minutes away from Fields’s house and the area where his crimes occurred. Adding to the similarities, Fields was charged with forcing his victims to enter his car at gunpoint, which was similar to the actions of the person who abducted Diane Hilliard. This court recognized that “Hilliard’s wife was the victim of a crime that was quite similar to the charges against Fields” when previously remanding this case for an evidentiary hearing. 309 F.3d 1095, 1105 (9th Cir.2002).
As the district court’s findings on remand demonstrate, Floyd Hilliard and his wife were far from oblivious to the similarity of the crimes at the time of Fields’s trial. When informed of the charges involved in the case at the beginning of jury selection, Floyd Hilliard doubted that he would be selected as a juror because of his wife’s parallel experience. In response to voir dire questioning about whether any close family members had been victims of a crime, Floyd Hilliard disclosed that his wife was “assaulted and beaten, robbed, two years ago.” The district court found that “he did not mention in voir dire that his wife had been raped or kidnapped” because he “did not want to be more explicit in open court than he was about what happened to his wife” and “thought the court and parties would understand that his use of the word 'assault’ in the context he did would encompass a sexual assault.” Fields v. Woodford, No. CV 92-0465 DT, slip op. at 44 (C.D.Cal. July 30, 2003). In other words, while the district court found Floyd Hilliard was not dishonest in his answer to the voir dire question, the emotional nature of the crime did consciously influence his answers, making him less forthcoming than he might otherwise have been.
After Floyd Hilliard was selected as a juror, his wife repeatedly suggested to him that Fields might be her rapist. In her deposition, Diane Hilliard recounted thinking that “there was a pretty good possibility” that Fields was responsible for her attack. Accordingly, she “begged” each night for his permission to attend the trial in order to see if Fields was her rapist. Floyd Hilliard, however, refused these repeated requests. He did, however, himself think during the trial about how the testimony “sound[ed] like what happened to my wife.”
The district court found that although Floyd Hilliard did not believe Fields was his wife’s abductor because Fields’s modus operandi had some differences from the details of her attack, he did not want her “to come to the trial because he did not want her to compromise him as a juror and he did not want the psychological trauma to affect their home life.... Mr. Hilliard was concerned his wife would be traumatized by the testimony and that it would create family problems.” Id. at 31. In his deposition, Floyd Hilliard also explained that he refused his wife’s request because “suppose indeed, in fact, she did ID him as the perpetrator ... then that would invalidate me as a objective juror.” He recognized such a development would have required him to disqualify himself as a juror and would have compromised the ongoing trial.
B.
Most jury bias claims- — “actual bias”— must be founded on evidence that a juror “was disposed to cast a vote against” the defendant. Dyer, 151 F.3d at 981. But “[i]n extraordinary eases, courts may pre*806sume bias based on the circumstances.” Id. In such cases, the doctrine of “implied bias”11 disqualifies the affected individual from serving on the jury, and dictates that any defendant whose fate was decided by such a juror has been denied his constitutional right to a fair trial by a panel of “impartial and indifferent” jurors. Morgan, 504 U.S. at 727, 112 S.Ct. 2222.
The doctrine of implied bias is premised largely on the understanding that certain circumstances create too great a risk of affecting a juror’s decisionmaking process, even if the juror is not, consciously, fully aware of the impact. The Supreme Court focused on this rationale in explaining why implied bias has traditionally disqualified individuals who had employment relationships with the parties from jury service:
Bias or prejudice is such an elusive condition of the mind that it is most difficult, if not impossible, to always recognize its existence, and it might exist in the mind of one (on account of his relations with one of the parties) who was quite positive that he had no bias, and said that he was perfectly able to decide the question wholly uninfluenced by anything but the evidence. The law therefore most wisely says that with regard to some of the relations which may exist between the juror and one of the parties, bias is implied, and evidence of its actual existence need not be given.
Crawford v. United States, 212 U.S. 183, 196, 29 S.Ct. 260, 53 L.Ed. 465 (1909); see also Smith v. Phillips, 455 U.S. 209, 221-22, 102 S.Ct. 940, 71 L.Ed.2d 78 (1982) (O’Connor, J., concurring) (“Determining whether a juror is biased or has prejudged a case is difficult, partly because the juror may have an interest in concealing his own bias and partly because the juror may be unaware of it.” (emphasis added)). In this respect, the implied bias doctrine applied to jurors parallels conflict-of-interest rules that apply to lawyers and judges. See Dyer, 151 F.3d at 983 n. 22. Those rules “ban[ ] lawyers and judges from taking on cases in certain conflict situations even if they are certain that the objective conflict will have no influence on them and are prepared to take every precaution to preclude such influence,” because “[hjuman self-perception regarding one’s own motives for particular actions in difficult circumstances is too faulty to be relied upon, even if the individual reporting is telling the truth as he perceives it.” United States v. Shwayder, 312 F.3d 1109, 1119 (9th Cir.2002). Floyd Hilliard’s own deposition attests to the wisdom of applying a similar rule to jurors: In explaining why he responded only “I doubt it,” rather than more definitively, to a voir dire question about whether it would be difficult for him to be an impartial juror, he noted “you can never be sure what’s in the back of your mind.”
Accordingly, the implied bias doctrine applies in circumstances where a juror “introduces ... [an] unpredictable factor into the jury room” or “introduces destructive uncertainties into the [factfinding] process.” Dyer, 151 F.3d at 982-83. In such cases, we do not consider the actual bias question of “whether [the juror] was disposed to cast a vote against [the defendant].” Id. at 981. It would not matter if the juror is found to have no actual bias against the defendant, because his “substantial emotional involvement” with some aspect of the case creates too great a risk of altering the jury’s deliberations despite *807the juror’s conviction that it will (or did) not. See Allsup, 566 F.2d at 71 (holding that implied bias applied to two bank tellers in a trial concerning the robbery of another branch of the bank for which they worked); see also United States v. Wood, 299 U.S. 123, 134, 57 S.Ct. 177, 81 L.Ed. 78 (1936) (describing implied bias as “a bias attributable in law to the prospective juror regardless of actual partiality”). As Judge Kozinski cogently observed on behalf of an en banc court:
Of course, a juror could be a witness or even a victim of the crime, perhaps a relative of one of the lawyers or the judge, and still be perfectly fair and objective. Yet we would be quite troubled if one of the jurors turned out to be the prosecutor’s brother because it is highly unlikely that an individual will remain impartial and objective when a blood relative has a stake in the outcome. Even if the putative juror swears up and down that it will not affect his judgment, we presume conclusively that he will not leave his kinship at the jury room door. The effect of this factor would be impossible to predict: Would the juror yield to his sympathies, or fight them and lean the other way? There is no way to know, but permitting such a juror to serve would introduce into the jury room an extraneous influence that could materially color the deliberations.
Dyer, 151 F.3d at 982. In other words, the implied bias doctrine exists principally to disqualify jurors who have an excess probability of being influenced in their deliberations by an extraneous consideration, despite their good faith belief that they can avoid doing so. It is not directed primarily at uncovering jurors hiding their conscious bias, although in some cases of implied bias that may be so.12
We have identified certain “general fact situations where bias might be presumed or implied.” Coughlin v. Tailhook Ass’n, 112 F.3d 1052, 1062 (9th Cir.1997). One of those situations is “where a juror or his close relatives have been personally involved in a situation involving a similar fact pattern.” Tinsley, 895 F.2d at 528 (emphasis added). Our inquiry, however, must be specifically tailored to the facts alleged to create implied bias, because we must “hesitate before formulating categories of relationships which bar jurors from serving.” Id. at 527.
In conducting this specifically tailored analysis, we must determine the effect on “an average person in the position of the juror in controversy,” because “the implied bias standard is essentially an objective one.” United States v. Gonzalez, 214 F.3d 1109, 1112-13 (9th Cir.2000) (emphases omitted) (quoting United States v. Cerra-to-Reyes, 176 F.3d 1253, 1260-61 (10th Cir.1999)).13 So, in evaluating whether the implied bias doctrine applies, we disregard a juror’s claims that he was not affected by his connection to the case. Id. at 1113. We do so because an underpinning of the *808implied bias doctrine is the recognition that the juror will often be unable to see for himself the effects of the connection. As Dyer observed, “[a juror] may declare that notwithstanding these prejudices he is determined to listen to the evidence, and be governed by it; but the law will not trust him.” 151 F.3d at 984 (quoting United States v. Burr, 25 F. Cas. 49, 50 (C.C.D.Va.1807) (No. 14,692g) (Marshall, J.)) (internal quotation marks omitted). This lack of trust, I emphasize once more, is not an unrebuttable presumption of dishonesty — or of lack of good faith — but, instead, a practical recognition of the complexities of human mental processes.
After engaging in the objective inquiry dictated by our precedents, I can only conclude that the deliberative process is likely to be affected for an average person who has both (1) feared for the safety of a spouse previously victimized by a quite similar, highly violent, terribly upsetting crime committed by a perpetrator still at large, and (2) been confronted with his spouse’s repeated speculations about a link between the defendant and her rapist every day of the trial, including requests to attend the trial in light of such speculation. The average person in such a situation undoubtedly will feel “substantial emotional involvement” with the crimes charged against the defendant, which we have noted is an indicia of implied bias. Tinsley, 895 F.2d at 527 (quoting Allsup, 566 F.2d at 71).
Indeed, there is direct evidence that Floyd Hilliard’s emotional involvement did influence his thought process to some degree during the trial: Floyd Hilliard stated during his 2003 deposition, it was “only a natural response” for him to recognize during the course of the trial that the testimony “sounds like what happened to my wife,” and the trial did “trigger a memory that ... my wife had a similar thing.” And Floyd Hilliard was concerned during the trial that the link between the facts developed during Fields’s trial and those of the crime against his wife were so close that it could be traumatic for his spouse to attend the trial. As Floyd Hilliard was far from indifferent during Fields’s trial to the similarities with the crime against his wife, it would have been an equally “natural response” for the similarities to affect his final deliberative process, without him being quite aware of the impact. Cf. Allsup, 566 F.2d at 71-72 (observing implied bias arose in part because the juror had a “reasonable apprehension of violence” from those who committed the crime on trial).
In short, permitting the average juror in Floyd Hilliard’s very unusual position to serve as a juror “would introduce into the jury room an extraneous influence that could materially color the deliberations. The juror in question would be lacking the quality of indifference which, along with impartiality, is the hallmark of an unbiased juror.” Dyer, 151 F.3d at 982. Accordingly, Fields’s constitutional right to a fair trial by a panel of “impartial and indifferent” jurors was violated by Floyd Hilliard’s presence in deliberations. See Morgan, 504 U.S. at 727, 112 S.Ct. 2222.
Given these close, emotionally charged, and ongoing links between Fields’s trial and the rape of Diane Hilliard in the same area a couple of years earlier by a rapist who was still at large, the majority’s observation that “[b]eing the spouse of a rape victim is not, in and of itself, such an ‘extreme’ or ‘extraordinary’ situation that it should automatically disqualify one from serving on a jury in a case that involves rape” is quite beside the point. Maj. op. at 774 (emphasis added). Our case law indicates that we should not inquire at such a general level about the effect of a relationship with the trial. See Tinsley, 895 F.2d at 527. That admonition applies in spades in this case. For, although the majority *809tries to segregate out and minimize the impact of the Hilliards’ struggle during the trial over whether Diane Hilliard could attend, in fact that struggle is critical in assessing the implied bias issue in this case.14 The daily discussions between Floyd Hilliard and his wife plainly had emotional content of their own and brought back to Floyd Hilliard on a daily basis the emotional impact of the connection between the brutal facts proved at trial and the brutal attack on his wife. No other case of which I am aware concerning implied bias arising from similarities between the offense being tried and the experiences of a juror’s close relative has this exceptional feature — that the juror could not put the link fully behind him during the trial, because it kept arising, daily, at home. And, as the jury deliberations took place at the end of the trial, not the beginning, the impact of the mid-trial discussion is not a side issue, as the majority supposes, but is all-important.15
C.
The majority, nonetheless, concludes that it “see[s] no basis for inferring bias now as a matter of law,” “[gjiven Hilliard’s honest response on voir dire that revealed a potentially disqualifying relationship, but not an extreme or extraordinary one, and the results of the evidentiary hearing which disclosed no actual bias.” Maj. op. at 775 (emphasis added). But neither Floyd Hilliard’s honest response nor his lack of actual bias answer the dispositive implied bias question — whether this case’s facts present an extraordinary situation to which implied bias applies.
1. In rejecting Fields’s implied bias claim, the majority stresses the district court’s finding that Floyd Hilliard was not dishonest during voir dire. See Maj. op. at 773 (“[D]ishonesty in voir dire is the critical factor.”); id. at 773 (“Although we have recognized that bias may be implied where close relatives of a juror ‘have been personally involved in a situation involving a similar fact pattern,’ we have never actually done so when the juror was honest on voir dire. We decline to do so here.” (citations omitted)). In my view, given (1) the close similarities between the accusations against Fields and the crime against *810Diane Hilliard combined with (2) the strong, fully understandable emotional reaction of the Hilliards to the crime and (3) the fact that Floyd Hilliard had to deal with his wife’s repeated suggestions during the trial that Fields was her assailant, Floyd Hilliard’s honesty or lack thereof matters little to the implied bias analysis.
As a general matter, the implied bias question is analytically distinct from the question of whether a juror was honest during voir dire: The former determines whether a juror was categorically unfit to serve as an impartial, indifferent juror; the latter determines whether a juror shortchanged the defendant’s right to learn about jurors’ proclivities through voir dire.16
The separate opinions by the Supreme Court in McDonough Power Equipment, Inc. v. Greenwood establish this distinction. Focusing on the harm that the appellant alleged arose “because counsel lacked an item of information which objectively he should have obtained from a juror on voir dire examination,” the principal opinion held that an appellant “must first demonstrate that a juror failed to answer honestly a material question on voir dire ” to receive a new trial on that basis. McDonough, 464 U.S. at 555-56, 104 S.Ct. 845. Five Justices, however, made clear that a juror could still be substantively biased, and unsuitable for jury service, regardless of whether his or her lies undermined the appellant’s procedural voir dire right. See id. at 556-57, 104 S.Ct. 845 (Blackmun, J., joined by Stevens and O’Connor, JJ., concurring) (“[Rjegardless of whether a juror’s answer is honest or dishonest, it remains within a trial court’s option, in determining whether a jury was biased, to order a post-trial hearing at which the movant has the opportunity to demonstrate ... in exceptional circumstances, that the facts are such that bias is to be inferred.”); id. at 558, 104 S.Ct. 845 (Brennan, J., joined by Marshall, J., concurring in judgment) (“[F]or a court to determine properly whether bias exists, it must consider at least two questions: are there any facts in the case suggesting that bias should be conclusively presumed; and, if not, is it more probable than not that the juror was actually biased against the litigant. Whether the juror answered a particular question on voir dire honestly or dishonestly, or whether an inaccurate answer was inadvertent or intention, are simply factors to be considered in this latter determination of actual bias.” (emphasis added)).
This interpretation of McDonough has been adopted by four other circuits. See Gonzales v. Thomas, 99 F.3d 978, 985-86 (10th Cir.1996) (holding a defendant who fails to show a juror dishonestly answered a question during voir dire, pursuant to McDonough’s principal opinion, still has “the normal avenue- of relief available to a party who is asserting that he did not have the benefit of an impartial jury,” including the use of implied bias (quoting McDonough, 464 U.S. at 556, 104 S.Ct. 845 (Blackmun, J., concurring)) (internal quotation marks omitted)); accord Zerka v. *811Green, 49 F.3d 1181, 1186 n. 7 (6th Cir.1995); Amirault v. Fair, 968 F.2d 1404, 1405-06 (1st Cir.1992) (per curiam); Cannon v. Lockhart, 850 F.2d 437, 440 (8th Cir.1988). The majority’s opinion initially recognizes this interpretation of McDonough by explaining there are three theories of juror bias — McDonough-style bias, actual bias, and implied bias — but later collapses these theories through its singular focus on whether Floyd Hilliard told the truth during voir dire. Compare Maj. op. at 766, with id. at 771-75. See also id. at 772 (noting “the existence of safeguards against actual bias ” and describing voir dire as “[t]he prime safeguard” (emphasis added) (quoting Tinsley, 895 F.2d at 527-28) (internal quotation mark omitted)).17
Moreover, logic dictates that a juror’s honesty in voir dire cannot affect a determination that his ties to the case present an “exceptional circumstance” making him categorically unfit to deliberate. Whether such a juror tells the truth or lies during voir dire does not change the likelihood that his deliberative process will be unconsciously affected by his emotional involvement in the ease. To borrow from Judge Konziski’s analysis in Dyer,
[u]nder the [majority’s] logic, reasonable jurists could hold that [Fields] would have been accorded due process even if he had been convicted by a jury comprised of the following twelve individuals: (1) the mother of ... the prosecutor, (2)[the prosecutor’s] former law partner, (3) [Los Angeles’s] Chief of Police, (4) the Grand Dragon of the [Los Angeles] KKK, (5) the sister of [the victim] who died in the shooting, (6) [the murder victimas] mother, (7) the victim of [Fields’s] prior [rape], (8)[Fields’s] ex-wife, (9) the District Attorney, (10) a[ ][Los Angeles] councilman running for re-election on a “tough-on-crime” platform, (11) [Fields’s] cellmate, and (12) [Hilliard’s] wife .. . — so long as they had all sworn they [were] fair.
151 F.3d at 985. Like the Dyer majority, I believe the presence of any of these individuals on Fields’s jury would have deprived him of a fair trial. And that would be the case equally whether they lied during voir dire about their disqualifying characteristic or were fully truthful. Dishonesty is central to a McDonough-style claim, and is of relevance to an actual bias contention because lying on voir dire can suggest an attempt to avoid disqualification and thereby to act on one’s bias. But dishonesty during voir dire has little to do in general with the concerns underlying the implied bias doctrine.
Moreover, Floyd Hilliard’s honesty during voir dire is, if anything, doubly irrelevant to the implied bias inquiry because critical factors supporting a implied bias finding in this case — perhaps the most important — arose after voir dire. No amount of voir dire questioning could have uncovered the fact that Diane Hilliard suggested repeatedly during trial that Fields might be her attacker and begged Floyd Hilliard to allow her to attend the trial, and that Floyd Hilliard refused to accede, in part because of fear of traumatizing his wife. *812Although the majority fails to consider whether these emotionally charged interactions after voir dire created implied bias, see Maj. op. at 773-74, 775-76 (rejecting Fields’s claims that developments after voir dire rendered Floyd Hilliard a biased juror, because the district court found Floyd Hilliard had no actual bias and did not believe his wife’s speculation), such bias can arise from, or be reinforced by, events that occur during the course of the trial, see Brooks v. Dretke, 418 F.3d 430, 431, 434-35 (5th Cir.2005) (finding implied bias when a juror was arrested during trial); Hunley v. Godinez, 975 F.2d 316, 320 (7th Cir.1992) (holding that implied bias arose when jurors were robbed during deliberations, and noting that a juror’s exposure to the biasing factor occurs after voir dire “may present an even more compelling reason for applying the presumption of bias”).
In short, Floyd Hilliard’s honesty in voir dire is entirely besides the point in applying the implied bias doctrine to the unique circumstances before us.
2. Based on the district court’s finding that Floyd Hilliard was not biased in fact, and our deferential standard of review in reviewing such findings, the majority quickly dismisses Fields’s actual bias claim. Maj. op. at 768. I agree that the district court’s findings dictate such a result. That conclusion should end the discussion of whether Floyd Hilliard was actually biased against Fields. See Dyer, 151 F.3d at 981 (explaining the distinction of the actual bias and implied bias questions).
The majority, however, regularly refers to, the district court’s finding of no actual bias throughout the following nineteen pages purportedly devoted to the implied bias question. See, e.g., Maj. op. at 774 (“Here, the evidentiary hearing showed no actual effect of his wife’s experience, or of their conversation, on Hilliard’s ability to be fair and impartial.”). Moreover, the majority entirely dismisses any claim of implied bias arising from facts occurring after voir dire on the basis that “the district court afforded Fields an opportunity to show that Hilliard was not a fair and impartial juror” — the test for actual bias— and “[h]e failed to do so.” Maj. op. at 773. But these findings concerning Floyd Hilliard’s actual bias are irrelevant to implied bias. See Gonzalez, 214 F.3d at 1113 (“[A] court will, where the objective facts require a determination of such bias, hold that a juror must be recused even where the juror affirmatively asserts (or even believes) that he or she can and will be impartial.”); Dyer, 151 F.3d at 984 (“[A juror] may declare that notwithstanding these [implied] prejudices he is determined to listen to the evidence, and be governed by it; but the law will not trust him.” (quoting Burr, 25 F. Cas. at 50) (internal quotation marks omitted)). The majority’s attempt to side step the true issue in this case does nothing to change my determination that this case presents an “exceptional circumstance” in which implied bias exists.
Because of the many similarities between the rape of Diane Hilliard — a crime that greatly affected both Floyd Hilliard and his wife — and the crime that Fields was on trial for committing, combined with Diane Hilliard’s suggestions that Fields might be her rapist and her repeated requests to attend the trial that raised concerns about her emotional well-being, the required objective inquiry should lead us to hold that Floyd Hilliard “introduee[d] ... [an] unpredictable factor into the jury room” and “introduce[d] destructive uncertainties into the [factfinding] process.” Dyer, 151 F.3d at 982-83. Floyd Hilliard may “swear [ ] up and down that it [did] not affect his judgment, [but] we presume *813conclusively that he will not leave [these influences] at the jury room door.” Id. at 982. As in Dyer, “[m]ore is at stake here than the rights of petitioner; ‘justice must satisfy the appearance of justice.’ ” Id. at 983 (quoting Offutt v. United States, 348 U.S. 11, 14, 75 S.Ct. 11, 99 L.Ed. 11 (1954)); see also id. at 981 (noting the possibility that the impliedly biased juror “could have harbored some empathy for criminal defendants ” (emphasis added)). Such an appearance was lacking regardless of Floyd Hilliard’s claims of conscious fairness — and even if in fact Floyd Hilliard was, as far as he was aware, not influenced by the confluence of circumstances giving rise to implied bias.
IV.
The State of California accused Stevie Lamar Fields of committing a series of heinous crimes and maintains he should be executed. The Constitution entitled Fields to have a set of impartial jurors make such a determination in accordance with the state’s penal statutes. Because the record raises substantial doubts whether Fields’s fate was decided by such a jury, the state may not execute him absent a retrial. I therefore dissent.
. I agree with the majority that Fields's jury misconduct and jury bias claims do not seek the retroactive application of law foreclosed by Teague v. Lane, 489 U.S. 288, 109 S.Ct. 1060, 103 L.Ed.2d 334 (1989).
. Those factors were:
(a) The circumstances of the crime of which the defendant was convicted in the present proceeding and the existence of any special circumstances found to be true....
(b) The presence or absence of criminal activity by the defendant which involved the use or attempted use of force or violence or the express or implied threat to use force or violence.
(c) Whether or not the offense was committed while the defendant was under the influence of extreme mental or emotional disturbance.
(d) Whether or not the victim was a participant in the defendant's homicidal conduct or consented to the homicidal act.
(e) Whether or not the offense was committed under circumstances which the defendant reasonably believed to be a moral justification or extenuation for his conduct.
(f) Whether or not the defendant acted under extreme duress or under the substantial domination of another person.
(g) Whether or not at the time of the offense the capacity of the defendant to appreciate the criminality of his conduct or to conform his conduct to the requirements of law was impaired as a result of mental disease or the affects [sic] of intoxication.
*793(h) The age of the defendant at the time of the crime.
(i) Whether or not the defendant was an accomplice to the offense and his participation in the commission of the offense was relatively minor.
(j)Any other circumstance which extenuates the gravity of the crime even though it is not a legal excuse for the crime.
Cal. Penal Code § 190.3 (1977).
. See supra note 2.
. I fail to understand the majority's claim that the jury's own decision to deliberate based upon the Bible is less violative of the constitutional requirement of narrowly channeled sentencing discretion or prohibition on undercutting the jury’s sense of responsibility for imposing death than a prosecutor urging the same effect. Maj. op. at 781. The ultimate substantive concern is how the jury actually deliberates, not who caused them to deliberate in an unconstitutional manner.
. In a closely related context, the Supreme Court has been emphatic that the Sixth Amendment guarantee to "the impartiality of any jury that will undertake capital sentencing” includes the guarantee that jurors will not simply vote to impose the death penalty for any murder. Morgan, 504 U.S. at 728, 112 S.Ct. 2222. As a result, a death sentence is unconstitutional if the jury includes “even one” juror who "will fail in good faith to consider the evidence of aggravating and mitigating circumstances as the instructions require him to do." Id. at 729, 112 S.Ct. 2222 (emphasis added); see also Wainwright v. Witt, 469 U.S. 412, 422, 105 S.Ct. 844, 83 L.Ed.2d 841 (1985) (allowing a potential juror to be removed for cause when his voir dire answers suggest that "he refuses to follow the statutory scheme” and instead "might vote for death under certain personal standards” (emphases omitted)).
. The majority says that McNair referred to the Bible extracts as "extrinsic evidence” only "as a statement of fact.” Maj Op. at 782 n.22. The Eleventh Circuit so stated in a section called the "Merits,” and after a lengthy paragraph setting out the standards applicable to assessing the impact of extrinsic evidence. See McNair, 416 F.3d at 1307-08. In context the “undisputed” comment is not simply a statement of the parties' positions but an indication that the point is so obvious as to be indisputable.
. The Fourth Circuit has held that a state trial court did not contravene clearly established Supreme Court case law by refusing to consider a claim that jurors engaged in misconduct by reading from a Bible during deliberations. Robinson v. Polk, 438 F.3d 350, 363-64 (4th Cir.), cert. denied, - U.S. -, 127 S.Ct. 514, 166 L.Ed.2d 383 (2006). The court noted, however, that "our answer could possibly be different on de novo review." Id. at 363.
. The majority finds no fault with the district court's application of Rule 606(b). See Maj. op. at 778.
. This court has previously refused to apply a . "strong[] presumption” of prejudice to an extrinsic information claim presented in a habeas case and instead specified that the standard Brecht analysis applies. Thompson, 74 F.3d at 1575 & n. 1; see also Pyles v. Johnson, 136 F.3d 986, 992-93 (5th Cir.1998) (rejecting thaf Mattox/Remmer, rather than Brecht, provides the correct standard to judge the prejudice of an extrinsic information claim on habeas review); Bibbins v. Dalsheim, 21 F.3d 13, 16 (2d Cir.1994) (per curiam) (same). I believe, however, that a presumption of prejudice may properly apply to an extrinsic information claim even under Brecht. See Thompson, 74 F.3d at 1577-82 (Reinhardt, J., dissenting); see also McNair, 416 F.3d at 1307 (applying the presumption to an extrinsic information claim presented in a habeas case: "the State bears the burden of rebutting the presumption by showing that the jurors' consideration of the extrinsic evidence was harmless to the defendant”). Moreover, Thompson is in some tension with our application of the Mattox presumption, by requiring the government to "show[] that there was no reasonable possibility that the [error] influenced the verdict,” in a habeas case involving communication between a juror and a third party. Caliendo v. Warden of Cal. Men’s Colony, 365 F.3d 691, 698 (9th Cir.2004). Nevertheless, I will assume that the Mattox/Remmer presumption does not apply on habeas review of extrinsic information claims.
. Because the jury misconduct renders Fields’s death sentence unconstitutional, it is unnecessary to consider whether he received ineffective assistance of counsel during the penalty phase of the trial. I note, however, that his lawyer's blatant deficiencies in investigating his life history and in preparing a mitigation case for the penalty phase of the trial — which "fell below minimum standards” according to the evidentiary referee appointed by the California Supreme Court, 51 Cal.3d 1063, 1068, 275 Cal.Rptr. 384, 800 P.2d 862 (1990) (internal quotation marks omitted)— make the majority’s reliance on the weight of the aggravating evidence quite troubling.
. I use the "implied bias” nomenclature because it is the term used in the applicable case law. I note, however, that the term somewhat misleadingly suggests that the doctrine rests on the implication of "actual bias” — that is, conscious prejudice or prejudgment — as a matter of law from certain facts. As I explain below, the doctrine actually focuses on a different kind of effect on the deliberative process.
. There also may be jurors who do have thoughts during deliberations that reflect the relationship to the facts of the case that gives rise to implied bias, but suppress the memory of their actual deliberative process sufficiently that, after the trial, they swear up and down— and believe — that there was no actual bias whatever. The implied bias doctrine ferrets out jurors affected by that psychological phenomenon as well.
. In addition to the Tenth Circuit’s decision in Cerrato-Reyes, upon which our holding in Gonzalez relies, three other circuits have specified that implied bias analysis is concerned with the objective effect of the fact situation on an “average person.” Johnson v. Luoma, 425 F.3d 318, 326 (6th Cir.2005), cert. denied, - U.S. -, 127 S.Ct. 58, 166 L.Ed.2d 54 (2006); United States v. Torres, 128 F.3d 38, 45 (2d Cir.1997); Person, 854 F.2d at 664.
. The majority maintains, Maj. Op. at 774-75 n.13, that a consideration of the discussions between Hilliard and his wife during trial would transform the implied bias inquiry from an objective analysis of the relationship between a juror and the trial into a subjective one. This is not so. Our analysis is still based on objectively observable facts that make it possible to infer bias as a matter of law with regard to any reasonable juror who finds himself in the same circumstances. Here, those circumstances include the mid-trial interactions between Hilliard and his wife.
. The majority suggests that an implied bias analysis does not allow consideration of circumstances that arose after voir dire, and that such circumstances can only be analyzed as "extrinsic influence" or “ex parte communications." See Maj. Op. at 775 n.14. True, many circumstances that arise after voir dire and that are relevant to implied bias will also involve inappropriate ex parte communications or extrinsic influence — for example, if a family or business relationship were to develop between a juror and the prosecutor or defendant after the trial had begun. But the fact that a circumstance also involves ex parte communication does not prevent a holding of implied bias where an "extraordinary” relationship between the juror and the case arises only after the trial begins. So, for example, if after voir dire but during trial a juror were the victim of a crime similar to the one being tried, the appropriate inquiry would still be for implied bias. Such an event would be entirely collateral to the trial, involving no one connected to it nor any of the facts before the jury, yet could influence an average person’s impartiality. Moreover, my analysis is simply an application of the existing case law on implied bias to the unusual facts in this case, and not a new rule subject to the constraints of Teague as the majority suggests. Maj. Op. at 775 n.14.
. Of course, if an,honest voir dire answer causes the defendant to become fully aware of the facts creating the implied bias, and he fails to request that the judge excuse the juror for cause, then he cannot ask for his conviction to be reversed on the jury bias ground. See United States v. Bolinger, 837 F.2d 436, 439 (11th Cir.1988) (per curiam) (“Thus, where the defendant or defense counsel knows of juror misconduct or bias before the verdict is returned but fails to share this knowledge with the court until after the verdict is announced, the misconduct may not be raised as a ground for a new trial.”). Because Floyd Hilliard’s voir dire answers did not — and could not, because many of the relevant facts arose only after jury selection— reveal all of the relevant facts, waiver does not apply in this case.
. I accept the majority’s implicit suggestion that implied bias operates to exclude a broader category of individuals when raised as a for-cause challenge during voir dire than when raised for the first time on appeal. See Maj. op. at 773 (suggesting follow-up questioning of Floyd Hilliard could have established implied bias mandating his excusal for cause); cf. Torres, 128 F.3d at 46-47 (noting the category of "inferable bias,” which unlike "implied bias” does not absolutely disqualify a juror but allows a trial judge to sustain a for-cause challenge without a showing of actual bias). I note, however, that the Supreme Court’s explanations of implied bias have come in cases — McDonough and Phillips — in which the juror bias claim was raised for the first time on appeal.