dissenting.
This is the first appellate case involving a religious hate crime under the Hate Crimes Prevention Act of 2009, 18 U.S.C. § 249. While I respect the majority’s efforts to construe a deceivingly simple, but actually complex, statute, I dissent. In my view, the majority has adopted an unduly restrictive interpretation of the statute.
Since this case was tried, the Supreme Court decided the case of Burrage v. United States, — U.S.-, 134 S.Ct. 881, 187 L.Ed.2d 715 (2014). The majority correctly holds that the “because of’ phrase used in § 249(a), similar to “results from,” requires proof that one act would not have happened “but for” the other. I disagree, however, with the majority’s conclusion that the trial court’s causation-instruction error was not harmless. This disagreement stems not from a dispute over the standards governing a harmless error analysis, but rather is from a disagreement over statutory construction.
First, although the majority correctly defines “but-for causality,” it ultimately fails to apply this correct definition to the facts of this case. As explained below, the pertinent inquiry is not whether religion was the sole but-for cause of the victims’ injury, but rather whether it was a but-for cause. Here, this requires that “bodily injury” would not have happened but for “the religion of the victim.” Overwhelming and uncontested evidence adduced at trial demonstrates that “but for” the victims’ Amish religion, their beards and hair would not have been cut. Because the record contains no evidence undermining the conclusion that the victims’ Amish religion was a but-for cause of the injury, the trial court’s causation-instruction error was harmless.
Second, although the majority acknowledges that the statute targets conduct and requires a but-for causal link between the bodily injury and the religion of the victims, it reads into the statute an extra, non-textual element. Specifically, the majority has effectively added to the hate crimes statute proof of faith-based animus, which is nowhere found in the statute. As discussed below, a plain language reading of the “because of’ provision in § 249(a) requires only a causal connection between the assailant’s conduct and the victim’s protected class. Thus, record evidence bearing on the existence (or absence) of this connection is of paramount importance to our harmless error analysis.
Finally, even following the majority’s view of statutory construction in this case, I conclude that the conviction against Samuel Mullet, Sr., must stand. The ringleader of the sixteen-member conspiracy, Bishop Samuel Mullet, Sr., confessed, repeatedly, to his religious motivation in orchestrating the assaults. The jury watched and heard him tell a television camera crew, “it’s all religion.” The jury received a news article in which Mullet said, “the goal of the haircutting was to send a message to the Amish in Holmes County.” He also said he “should be allowed to punish people who break the laws of the church.” Mullet never denied these statements. He produced no evidence to the contrary. The majority’s finding that a jury could have found some other “but for” cause of the assaults, in the absence of any record evidence supporting such a competing cause, is not reasonable. Thus, at a minimum, the trial court’s causation-instruction error was harmless as to Samuel Mullet, Sr.
I.
A. But-For Causality
As the majority acknowledges, Burrage, decided after the trial in this case, provides definitive guidance concerning the *604appropriate construction of the term “because of’ that the trial court simply did not have at the time of trial. At issue in Burrage was a statutory enhancement provision that calls for increased penalties “on a defendant who unlawfully distributes a Schedule I or II drug, when ‘death or serious bodily injury results from the use of such substance.’ ” 134 S.Ct. at 885 (emphasis added) (quoting 21 U.S.C. § 841(a)(1), (b)(l)(A)-(C) (2012)). The purchaser of the drug died following a binge consisting of five drugs, including heroin, which the defendant had distributed. Id. Both of the medical experts testifying at trial opined that the heroin was a contributing factor, but neither could say whether the decedent would have lived had he not taken the heroin. See id. at 885-86. The parties disputed the appropriate language for the jury instruction regarding causation. See id. at 886.
It is within this context that the Burrage Court set about defining the phrase “results from.” Within its analysis, the Court reviewed both case law and dictionary definitions of the phrase “because of,” noting that the definitions of the phrases “because of’ and “results from” “resemble” one another, in that each requires but-for causality. Id. at 889. The Burrage Court concluded that these phrases “require[] proof ‘that the harm would not have occurred in the absence of—that is, but for— the defendant’s conduct.’ ” Id. at 887-88 (quoting Univ. of Tex. Sw. Med. Ctr. v. Nassar, — U.S. -, 133 S.Ct. 2517, 2525, 186 L.Ed.2d 503 (2013) (citation omitted) (internal quotation marks omitted)).
To better explain what but-for causality means, the Burrage Court included a number of illustrative examples:
[W]here A shoots B, who is hit and dies, we can say that A [actually] caused B’s death, since but for A’s conduct B would not have died. The same conclusion follows if the predicate act combines with other factors to produce the result, so long as the other factors alone would not have done so—if, so to speak, it was the straw that broke the camel’s back. Thus, if poison is administered to a man debilitated by multiple diseases, it is a but-for cause of his death even if those diseases played a part in his demise, so long as, without the incremental effect of the poison, he would have lived.
This but-for requirement is part of the common understanding of cause. Consider a baseball game in which the visiting team’s leadoff batter hits a home run in the top of the first inning. If the visiting team goes on to win by a score of 1 to 0, every person competent in the English language and familiar with the American pastime would agree that the victory resulted from the home run. This is so because it is natural to say that one event is the outcome or consequence of another when the former would not have occurred but for the latter. It is beside the point that the victory also resulted from a host of other necessary causes, such as skillful pitching, the coach’s decision to put the lea-doff batter in the lineup, and the league’s decision to schedule the game. By contrast, it makes little sense to say that an event resulted from or was the outcome of some earlier action if the action merely played a nonessential contributing role in producing the event. If the visiting team wound up winning 5 to 2 rather than 1 to 0, one would be surprised to read in the sports page that the victory resulted from the leadoff batter’s early, non-dispositive home run.
Id. at 888 (citations omitted) (internal quotation marks omitted); see id. at 892 (Ginsburg, J., concurring) (“I do not read *605‘because of in the context of antidiscrimi-nation laws to mean ‘solely because of.’ ”).
What these examples demonstrate is that there often exists more than one but-for cause; a number of necessary causes may operate concurrently to produce a given outcome.1 Such is the case here. The majority submits that the evidence supports an additional but-for cause of the hair sheering, namely, interpersonal and intra-family disagreements. But contrary to the majority’s suggestion, neither this Court nor a jury needs to “untangl[e] the role[s]” these various factors played. Supra at 595. Nor does it matter whether these other factors were a more significant but for cause or whether one of them played a “starring role” at trial. Id. at 594; see Burrage, 134 S.Ct. at 888 (the poison was a but-for cause even though it had only an “incremental effect”). Rather, as the Burrage Court made clear through its examples, regardless of whether other causes are necessary to the outcome, the pertinent inquiry is always this: in the absence of the cause or factor at issue, would the statutorily prohibited outcome have occurred? As applied in Burrage, the Court concluded that the defendant could not be convicted under the enhancement statute because the Government conceded that there was no evidence that the decedent would have lived but for (or in the absence of) his heroin intake. Id. at 892.
Under § 249(a), the factor at issue is the victims’ protected class—here, the victims’ “actual or perceived [Amish] religion.” 18 U.S.C. § 249(a)(2). The pertinent but-for causality inquiry, then, is whether, even if all of the other contributing or but-for factors remained, the prohibited conduct (the beard and hair cutting) would have occurred but for or in the absence of the victims’ Amish religion. In more concrete terms, would Defendants have cut the victims’ hair and beards if the victims were Catholic, atheist, or any other non-Amish faith?
Although the majority recognizes that the phrase “because of’ “indicates a but-for causal link between the action that comes before it and the circumstance that comes afterwards,” supra at 591 (emphasis added), the majority endorses a different interpretation of “because of’ within its harmless error analysis. In effect, the majority construes the “because of’ provision of § 249(a) to require the victim’s protected class to be the but-for cause rather than a but-for cause. This runs afoul of Burrage. See id. at 591-92, 594.
B. The Statutory Elements to be Linked
In its harmless error analysis, the majority reasons that “assaults involving religious symbolism do not invariably stem from religious motives.” Id. at 595. This language reflects that the majority in effect interprets § 249(a) to require an additional, non-textual element, namely, a showing of faith-based animus. But the Supreme Court advises against construing statutes to include non-textual elements. See, e.g., Nassar, 133 S.Ct. at 2528-29 (“[I]t would be improper to conclude that what Congress omitted from the statute is nevertheless within its scope.” (citing Gardner v. Collins, 27 U.S. 58, 2 Pet. 58, *60693, 7 L.Ed. 347 (1829) (“What the legislative intention was, can be derived only from the words they have used; and we cannot speculate beyond the reasonable import of these words.”))). Instead, “[t]he role of this Court is to apply the statute as it is written-even if we think some other approach might ‘accor[d] good policy.’” Burrage, 134 S.Ct. at 892 (second alteration in original) (quoting Comm’r v. Lundy, 516 U.S. 235, 252, 116 S.Ct. 647, 133 L.Ed.2d 611 (1996)). Nothing in the statute requires faith-based animus.
As the majority points out, a plain-language reading of the “because of’ provision in § 249(a) connotes a “but-for causal link between the action that comes before it and the circumstance that comes after-wards.” Supra at 591. And as the majority further acknowledges, § 249(a) “targets conduct, not bigoted beliefs,” id. at 592, requiring a causal connection between the assailant’s conduct and the victim’s protected class, not a causal connection between the assailant’s bigoted beliefs and the victims protected class. The state-court decisions that the majority flags as having construed similar state hate crime statutes agree. See In re M.S. 42 Cal.Rptr.2d 355, 896 P.2d at 1375 (“ ‘Because of ... connotes a causal link between the victim’s characteristic and the offender’s conduct....”); Plowman, 838 P.2d at 561 (stating that the statute “expressly and unambiguously requires the state to prove a causal connection between the infliction of injury and the assailants’ perception of the group to which the victim belongs”). By way of contrast, the text of the penalty enhancement statute at issue in Wisconsin v. Mitchell does include an additional element, calling for increased penalties where the assailant “[ijntentionally selects the person against whom the crime ... is committed ... because of the [victim’s membership in the protected class].” 508 U.S. 476, 480, 113 S.Ct. 2194, 124 L.Ed.2d 436 (1996) (emphasis added). Thus, the explicit text of the enhancement statute at issue in Mitchell requires an additional but-for causal link between the assailant’s selection of the victim and the victim’s protected class. The text of § 249(a), unlike the provision in Mitchell, requires only a causal connection between the assailant’s infliction of the injury (beard and hair cutting) and the victims’ actual or perceived protected class (Amish religion). To this, the effect of the majority’s construction is to incorrectly add a requirement wholly absent from the statute—that the defendants must also act with faith-based animus.
C. Application of the Harmless Error Analysis
Applying § 249(a) as written and defining but-for causality in accordance with Burrage requires the conclusion that the trial court’s instruction of the jury on the issue of causation was harmless. As the majority acknowledges, in the event of an erroneous jury instruction, the reviewing court analyzes whether the error is harmless, specifically asking whether it is “beyond a reasonable doubt that the omitted element was uncontested and supported by overwhelming evidence, such that the jury verdict would have been the same absent the error.” Neder, 527 U.S. at 17, 119 S.Ct. 1827. The question, in other words, is “whether the record contains evidence that could rationally lead to a contrary finding with respect to the [element in error].” Id. at 19, 119 S.Ct. 1827. As detailed above, the proper but-for causality inquiry is whether, even if all of the other contributing or but-for factors remained, the prohibited conduct (the beard and hair cutting) would have occurred but for or in the absence of the victims’ Amish religion.
The uncontested and overwhelming evidence adduced at trial demonstrates that Defendants chose to inflict the beard and *607hair-cutting injuries upon the victims precisely because of the victims’ religion.2 The testimony at trial demonstrates that the parties disagreed about the doctrinal significance of the beard and about whether the cutting of the beard (and hair) was acceptable. Defendants knew that their conduct would inflict great harm because of the victims’ religious beliefs on these issues. On this point, Dr. Kraybill, an expert in the Amish faith, testified that that the Amish world views men’s beards as “sanctioned by God,” R. 541 at 8, as a “sacred tradition,” id., “a symbol of devotion to God, a symbol of piety, a symbol of righteousness, and a very public symbol of Amish religious identity,” and something “[thatj’s there all the time” as a “central” piece “of a man’s religious identity,” id. at 8-9. Dr. Kraybill further testified that, as a practice “woven into Amish religious tradition,” Amish “women don’t cut their hair and keep their hair long.” Id. at 10. No evidence to the contrary was presented.
In addition to Dr. Kraybill’s unrefuted testimony, other evidence adduced at trial establishes the requisite causal link between the beard and hair cuttings and the victims’ Amish faith. See, e.g., R. 556-11 at 8 (Mullet’s statements to his son, Johnny Mullet, indicating that cutting off hair and beards is “a religious degrading”); R. 556-10 at 6 (same as to Levi Miller); R. 540 at 11 (Barbara Yoder testifying that the beard and hair cuttings “would help stop people” from “being ... Amish hypocrites”); R. 529 at 85 (Nancy Burkholder testifying that she and her brothers cut their parents’ hair “to help them ... liv[e] a more proper Amish life”).
The majority identifies no evidence sufficient to support a contrary finding. As discussed above, the interpersonal and in-tra-family disagreements the majority identifies lend support to the conclusion that there may exist more than one but-for cause in provoking the assaults. But, as the Supreme Court recently held, the existence or demonstration “of other necessary causes” “is beside the point.” Burrage, 134 S.Ct. at 888. Instead, for purposes of our harmless error analysis, what matters is whether Defendants have offered evidence to support the 'conclusion that they would have cut the victims’ hair and beards even if the victims were not Amish. There exists no such evidence, let alone evidence that “could rationally lead to a contrary finding.” Neder, 527 U.S. at 19, 119 S.Ct. 1827. Accordingly, because the requisite causal link between the beard and hair cuttings and the victims’ Amish faith is supported beyond a reasonable doubt, I conclude that the trial court’s causation-instruction error was harmless such that all of the convictions must stand.
II.
Even accepting arguendo the majority’s view of causation, the conviction against Samuel Mullet, Sr. must stand.3 Over*608whelming evidence from trial, including Mullet’s confession that the disputes were “all religion,” demonstrates that Mullet participated in the assaults because of the victims’ religious beliefs. Mullet put forth no evidence at trial that he participated in the attacks for any other reason than religion. This requires the conclusion that the trial court’s causation-instruction error was harmless as to Mullet.
Harrington v. California, 395 U.S. 250, 89 S.Ct. 1726, 23 L.Ed.2d 284 (1969), is instructive. In Harrington, the Supreme Court found that the trial court’s admission of confessions from petitioner’s codefendants who did not take the stand violated his Confrontation Clause rights according to Bruton v. United States, 391 U.S. 123, 88 S.Ct. 1620, 20 L.Ed.2d 476 (1968). 395 U.S. at 252, 89 S.Ct. 1726. But the Harrington Court nevertheless affirmed the conviction because it found the error harmless. Id. at 254, 89 S.Ct. 1726. The Court explained that the petitioner all but confessed to the crime by “ma[king] statements” that “placed him at the scene of the crime” and that the additional evidence of his guilt was “so overwhelming” that to reach a contrary conclusion would be tantamount to holding “that no violation of Bruton can constitute harmless error.” Id. at 252-54, 89 S.Ct. 1726. Where, as here, the record reflects a confession and overwhelming evidence of guilt, the Sixth Circuit has, until now, consistently followed Harrington’s lead. See, e.g., Hartford v. Haskell, Nos. 90-2017, 90-2122, 1991 WL 169333, at *2 (6th Cir. Sept. 4, 1991); Burks v. Perini, No. 85-3507, 1986 WL 18388, at *1 (6th Cir. Nov. 25, 1986). These cases demonstrate that where the record reflects a confession and other overwhelming evidence of guilt, the conviction must stand. Such is the case here: the jury verdict would have been the same absent the trial court’s causation-instruction error.
Mullet first confessed his motivation for the assaults in October 2011 when he spoke to media outlets about the attacks. The jury saw a clip of Mullet’s interview on a local-news broadcast in which he stated, “it’s all religion” and explained that the conflicts “started with us excommunicating members” of Bergholz. Mullet also gave an interview to the Associated Press (AP) around the same time. The jury heard the following passages from the AP story:
The leader of an Amish group with members accused of going into the home of another Amish man and cutting his hair and beard said Monday, [“]It’s a religious matter and police shouldn’t be involved.[”]
Mullet, 66, said the goal of the haircut-ting was to send a message to Amish in Holmes County that they should be ashamed of themselves for the way they were treating Mullet and his community. [“]They changed the rulings of our church here, and they’re trying to force their way down our throat, make us do like they want us to do.[”]
Mullet said he should be allowed to punish people who break the laws of the church, just as police are allowed to punish people who break the laws of the state. [“]You have your laws on the road and the town[;] if somebody doesn’t obey them, you punish them. But I’m *609not allowed to punish the church people[?”] Mullet said. [“]I just let them run over me? If every family would just do as they pleased, what kind of church would we have [?”]
R. 540 at 143-45 (emphasis added).
The jury also heard Mullet confess his motivation for the assaults yet again within the context of jailhouse phone calls with his co-defendants. In one such conversation, Mullet, speaking to Levi Miller, said, “Well listen[,] Levi[,] you only cut off hair----Public is going to say [’]wow taking somebod[y’s] hair off[,] what is that, that is a religion])] Fred [Abdallah (the Sheriff) ] said it right in the paper [that] it is a religious degrading.” R. 556-10 at 6. And, in a similar conversation with Johnny Mullet, Samuel Mullet, Sr. said, “To take off hair[,] Fred had it right in the paper[:] it’s a religious degrading to cut the hair and the beard.” R. 556-11 at 8.
Moreover, the jury heard testimony from Barbara Yoder, one of Mullet’s daughters, and Daniel Shrock, one of his grandsons, corroborating Mullet’s stated motivation. Specifically, Barbara Yoder testified that Mullet believed the victims were “Amish hypocrites.” R. 540 at 11-13. When asked what her “dad was saying about beard and haircuttings in the ... fall of 2011,” Barbara answered that Mullet said the hair cuttings “would help stop people” from “being ... Amish hypocrites.” Id. at 11. Daniel Shrock similarly testified that Mullet spoke about the Millers’ assault before it happened, saying that it “might help” them live a more “spiritual life.” R. 537 at 190.
In addition to this direct evidence, the jury heard circumstantial evidence supporting its conclusion that Mullet took part in the assaults because of the victims’ religion. For example, it heard that the Millers and Shrocks, victims of the assaults, left or were excommunicated from Ber-gholz because of religious differences with Mullet. It heard that another victim, Raymond Hershberger, was on the committee that voted not to follow Mullet’s excommunications, which, as Mullet said himself, undermined his authority to punish people who broke the rules of his church, see R. 540 at 145. It also heard Myron Miller, another victim in this case, explain that Mullet asked him to leave a meeting because of a disagreement about “church matters,” R. 537 at 16; and that Miller’s church refused to take communion with Mullet’s, see id. at 22, based on its concerns regarding “cultic activities” at Ber-gholz, id. at 85.
The record does not, however, contain any evidence upon which a jury could have rationally relied to conclude that religion did not motivate Mullet to participate in the assaults. Thus, given Mullet’s unequivocal confession of his religious motive and the additional, unrefuted evidence corroborating his confessed motive, the erroneous causation instruction was harmless as to Mullet. See Neder, 527 U.S. at 17-19, 119 S.Ct. 1827; see also Harrington, 395 U.S. at 252-54, 89 S.Ct. 1726; Haskell, 1991 WL 169333, at *2; Perini, 1986 WL 18388, at *1.
The majority reaches a different conclusion with regard to Mullet. In their view, a jury “might reasonably conclude” that Mullet shared “similar [allegedly] non-faith-based motives” with the other assailants, supra at 599, because he “had close personal relationships with the[m],” id. at 599. The majority believes that the nature of a conspiracy conviction also supports this theory. Alternatively, they submit that some evidence suggests that Mullet “agreed to the assaults because he felt the victims had disrespected him, and he wanted revenge for that disrespect.” Id. at 599.
*610Neither the record nor the governing authority offers support for the majority’s theories. The evidence the majority cites supporting the theory that Mullet acted because he felt disrespected in no way undercuts Mullet’s confessed motive for the attacks. For example, the majority cites Mullet’s statement that “the goal of the haircutting was to send a message to [the] Amish in Holmes County.” R. 540 at 143. Yet the very next sentence in his own statement confirms the religious nature of Mullet’s motive: “They changed the rulings of our church here, and they’re trying to force their way down our throat, [to] make us do like they want us to do.” Id. at 143-44. The majority’s reliance upon Mullet’s statements to an FBI agent that some of the other assailants were “mad” about the way the victims had treated him is equally unavailing. Id. at 123. The unequivocal evidence demonstrates Mullet’s religious motives. There is no evidence of record connecting Mullet’s speculation regarding the other assailants’ motivation to his own. Finally, the majority’s reliance upon testimony from Mullet’s sister that he had a bad temper similarly lacks relevance given that the record reflects that no one offered testimony connecting Mullet’s bad temper to the assaults. As the Neder Court explained:
A reviewing court making [a] harmless-error inquiry does not ... become in effect a second jury to determine whether the defendant is guilty____ Rather a court, in typical appellate-court fashion, asks whether the record contains evidence that could rationally lead to a contrary finding with respect to the omitted element.
527 U.S. at 19, 119 S.Ct. 1827 (citation omitted) (internal quotation marks omitted).
For this same reason, the majority’s alternative theory that a jury “might reasonably conclude” that Mullet shared the same non-religious motivations as the other assailants is simply not viable. During closing arguments, Mullet’s counsel hinted that family or community infighting might have motivated Mullet. See R. 542 at 207, 214. But neither a jury nor this Court may speculate as to a theory raised only in closing arguments that is not supported by the evidence. And although a jury may draw an inference from established facts, no such facts exist here. No witnesses testified that Mullet had other reasons for the assaults. The defense adduced no evidence on cross examination that could support even an inference that Mullet’s motivation differed from the religious motivation he explicitly and repeatedly endorsed. See Neder, 527 U.S. at 19, 119 S.Ct. 1827 (to demonstrate that a trial error was not harmless, the defendant must both “contest! ] the omitted element” and “raise [ ] evidence sufficient to support a contrary finding” (emphasis added)).
Finally, Mullet’s conspiracy conviction does not compel the conclusion that he acted with the same motivations as the other assailants. A conspiracy charge and a conspiracy conviction apply to each defendant individually. See, e.g., United States v. Pearce, 912 F.2d 159, 161 (6th Cir.1990) (“[T]he government must prove ... that each conspirator knew of, intended to join and participated in the conspiracy.” (emphasis added)). Thus, conspiracy is not an all-or-nothing conviction—some defendants can be acquitted while others can be convicted. See, e.g., United States v. Gibbs, 182 F.3d 408, 423 (6th Cir.1999). Within the context of this case, this means that evidence and theories that apply to some defendants do not automatically apply to Samuel Mullet simply because he knew those other defendants. See, e.g., id. (“[E]vidence that the defendants knew each other ... fails to prove membership *611in the conspiracy.”). In short, in order to benefit from his co-conspirators’ evidence, Mullet must actually connect it to himself. See id.; Pearce, 912 F.2d at 161. No such evidence came before the jury. Instead, all of the evidence adduced at trial—both direct and circumstantial—demonstrates that Mullet participated in the assaults because of the victims’ religious beliefs.
The overwhelming and unrefuted evidence adduced at trial demonstrates that Mullet participated in the assaults because of the victims’ religious beliefs. The record contains no “evidence that could rationally lead to a contrary finding.” Neder, 527 U.S. at 19, 119 S.Ct. 1827. Thus, even accepting the majority’s view of causation, I find that the trial court’s causation-instruction error was harmless as to Mullet such that his conviction must stand.
. Indeed, to further drive home this point, in quoting two other recent Supreme Court decisions analyzing the meaning of “because of,” the Burrage Court replaced those decisions’ utilization of the phrase "the but-for cause,” Nassar, 133 S.Ct. at 2528; Gross, 557 U.S. at 176, 129 S.Ct. 2343 (emphasis added), with the phrase "'[a] but-for cause,’” Burrage, 134 S.Ct. at 889 (emphasis added) (alterations in original) (quoting Nassar, 133 S.Ct. at 2528; Gross, 557 U.S. at 176, 129 S.Ct. 2343).
. The majority invokes an example of a neck injury from the forceful removal of a cross pendant after an unrelated dispute, and an example involving a child who tells his parents that he is gay. See supra at 596-97. Hypotheticals are helpful only to the extent they are analogous to the case at hand. Here, the jury found that sixteen defendants led by an Amish bishop angry at a decision made by other Amish clergy, acted in a violent manner to obstruct religious practices of the multiple victims in this case. There is no comparison between the majority's hypotheticals and the facts of this case.
. According to the majority, the Government did not “present a separate harmless-error argument as to [Samuel Mullet].” Supra at 599. The Government devoted a section of its brief to its position that the district court gave the correct causation instruction and, even if not, its "error is subject to the harmless-error analysis.” Appellee Br. at 102 (citing Neder, 527 U.S. at 9-10, 15, 119 S.Ct. 1827). Within *608this section, the Government argued that "[t]here is overwhelming evidence that the attacks took place because Mullet and his followers did not like that the victims disagreed with Mullet's religious practices.” Id. at 112. The Government then tied each of the attacks at issue back to discrepancies with Mullet over religion. See, e.g., id. I view this as more than sufficient for this Court to consider the harmless error issue as it applies to Mullet.