United States v. Rejon Taylor

814 F.3d 340 | 6th Cir. | 2016

Lead Opinion

ROGERS, J., delivered the opinion of the court in which McKEAGUE, J., joined, and WHITE, J., joined in part. WHITE, J. (pp. 379-98), delivered a separate opinion concurring in part and dissenting in part.

*345OPINION

ROGERS, Circuit Judge.

Rejón Taylor appeals his convictions and death sentence for carjacking resulting in death, kidnapping resulting in death, and using a firearm to commit murder while committing carjacking and kidnapping. His appeal raises fifteen issues, none of which warrants reversal or remand.

The district court described the crimes underlying this case as follows:

[Taylor] had been responsible for various thefts and burglaries from [Guy] Luck’s house and other nearby residences in Atlanta between 2001 and 2003. On August 6, 2003, [Taylor], along with codefendants Sir Jack Matthews and Joey Marshall, went to Luck’s house with the intention of robbing him. After confronting Luck at gunpoint, Marshall guarded Luck while [Taylor] began looking through Luck’s house. Inside the house, [Taylor] took around $600 or $800. Marshall testified [Taylor] later told him there was a warrant or other document connected with [Taylor’s] arrest on theft charges in another case, which suggested Luck could be a witness against [Taylor],
At gunpoint, Luck was forced outside his house and into his van. [Taylor] got in the driver’s seat, while Matthews guarded Luck in the back. [Taylor] and Matthews each had a gun. [Taylor] drove the van onto Interstate 75 and traveled north from Atlanta. They made a brief stop at a gas station in north Georgia before eventually crossing into southeast Tennessee, where [Taylor] exited the expressway and drove into the Chattanooga suburb of College-dale. During the trip, Marshall followed behind in a car registered to [Taylor’s] mother.
As [Taylor] drove the van around relatively isolated roads in Collegedale, there was a confrontation in the back of the van, in which Matthews fired a shot, which hit Luck in the arm. [Taylor] turned around from the driver’s seat and fired three shots at Luck. The third bullet hit Luck in the mouth and caused his death later that day at Erlanger Hospital. [Taylor] and Matthews left their guns in the van and walked briskly from the van to the car driven by Marshall. They then drove back to Atlanta.
[Taylor] was subsequently arrested and incarcerated pre-trial at the Hamilton County Jail in Chattanooga. While incarcerated there, [he] was part of a group of inmates that attempted to escape.

United States v. Taylor, 583 F.Supp.2d 923, 926-27 (E.D.Tenn.2008).

A jury convicted Taylor of carjacking resulting in death, kidnapping resulting in death, and using a firearm to commit murder while committing carjacking and kidnapping. The jury recommended a death sentence, which the district court imposed. The question in this case is whether particular acts or omissions in the course of Taylor’s trial require remand. Taylor has identified fifteen potential grounds for reversal or remand.

Ground I: Juror Bias

Taylor first argues that the district court should have more strenuously questioned Juror 1 (later determined to be the foreperson) to determine whether exposure to an out-of-court remark affected her ability to be impartial. Taylor claims that the district court’s failure to do so — or to replace Juror 1 with an alternate juror— requires remand for resentencing. While it might well have been advisable for the district court to ask more questions of Juror 1, the district court did not abuse its discretion in declining to press on with the questioning once the district court satisfied *346itself that Juror 1 had not been prejudiced by the out-of-court remark.

The jury found Taylor guilty on September 8, 2008. Just over a week later, on the eve of the sentencing hearing, the Government announced its intention to use at the hearing recordings of phone calls Taylor made from jail after the verdict. In one of those calls, the Government indicated, Taylor allegedly referred to the jurors as “racist rednecks.”1 Local media reported the alleged remark. On September 19, 2008, Taylor’s counsel moved for a mistrial and asked that the district court question the jurors outside the presence of counsel about their exposure to media reports of Taylor’s reported remark. Taylor’s counsel expressly declined the opportunity to participate in or be present at the questioning, theorizing that the jurors might be less forthcoming in his presence. The Government agreed that questioning by the court was appropriate and so, on September 23, 2008, the district court questioned each of the jurors and alternates privately, without the parties or their lawyers being present. The district court told the jurors that, because of the break between the guilt phase and the sentencing phase, the attorneys had asked the court to make sure that none of the jurors had been exposed to publicity about the trial. Eleven of the eighteen jurors and alternates reported that they had been exposed to publicity about the trial. The district, court asked all but one of the jurors who had heard about a “racist” or “redneck” remark whether the remark would affect their decision and whether they could put the remark out of their minds. All of the jurors who were asked assured the district court that the remark would not affect their ability to be impartial.

The district court questioned Juror 1 as follows:

THE COURT: The lawyers, just as a matter of precaution more than anything else, asked if I would meet with each one of the jurors this morning to make sure that no one was exposed to any publicity during the break. The break was pretty long. And I did not see any harm in acquiescing in their request, so I agreed to do so. During the break were you exposed to any publicity at all about the trial?
JUROR 1: Well, I’ll have to say that the only thing that really — and I did not read or anything — just that I’ve had a couple people say, “I didn’t know you were a redneck.” So, you know, obviously there are some comments that have been made along those lines. But, no, sir, I have not read anything or seen anything. And the only reason I knew we were going to talk today is because I turned the TV on, after turning off Dancing with the Stars last night, it was on Channel 3, and it said, “Judge Collier is going to talk — ” I couldn’t get it off fast enough. So ...
THE COURT: You did not follow up on the conversations, then, when people said they did not know you were a redneck?
JUROR 1: No. I obviously have speculated. In your head ... THE COURT: Did they tell you where that comment came from?
JUROR 1: Not necessarily, no, sir. I assumed it was from either the defendant or some of his friends. My colleague said to me last Tuesday — we were there Wednesday — “We knew you’d be here today.”
*347I said, “You did?”
So they knew about it, but ...
I said, “I can’t talk about it.”
THE COURT: Well, we are working when you guys are not here. We’re still doing things.
JUROR 1: I understand that. - But evidently they knew we weren’t. .
THE COURT: Okay. Well, thank you.
JUROR 1: Oh, a couple of things I’d like to ask if I may.
THE COURT: Uh-huh. (Moving head up and down.)
JUROR 1: Is it the usual procedure that the defendant knows the jurors’ names? I’ve had some concern, you know. It is a murder trial, after all. And there might be some repercussions down the pike, of being found and something happening, with friends. Is that normal for jurors’ names to be called? I know polling the jury, but is there a reason it couldn’t be Juror 1, Juror 2, that kind of thing?
THE COURT: We collect the information once the jury is impaneled. We give a long list, including information, and we have a standing — is it a rule, or just an order? — a practice that we collect the information. We don’t—
THE CLERK: I’m not sure if it’s an order. I know Mari puts a cover sheet on the jury list that it’s to be returned.
THE COURT: To be returned as soon as the jury is selected. The lawyers might have some of their own notes with names on it, but defendants would not have that.
JUROR 1: Okay.
THE COURT: They would have to have an extremely good memory to recall all of that. And we don’t put addresses on anything. So even if someone remembered a name, it would take quite an effort to go further than that.
JUROR 1: Well, with the Internet nowadays, it’s not so tough. So, you know, there was some concern. My husband and I, that’s the only thing we’ve talked about. So, thinking about getting a weapon.
I didn’t bring my purse. Okay.

After reviewing the jurors’ and alternates’ responses, Taylor moved for a mistrial due to incurable juror prejudice. In the alternative, Taylor requested a hearing during which counsel could question the jurors and alternates and after which the district court could decide whether to dismiss the entire jury or remove the most “taint[ed]” jurors. The district court heard argument on Taylor’s requests and ultimately denied them both. Noting “the obvious risk” of reinforcing the allegedly prejudicial publicity by further questioning jurors and alternates about it, the district court observed that a simple stipulation— that Taylor “never said [racist rednecks]” — could resolve the issue without running the risk of further juror prejudice. In a memorandum and order elaborating on its ruling, the district court stated:

In this case, the Court had the benefit of personally interviewing the jurors. From these interviews, it was evident that those who were aware of the comment were not nearly as impacted by the comment as one might imagine. Most of the jurors who were aware of the comments heard them from co-workers who joked that they were rednecks. Most of those jurors did not know what the jokes referred to and many did not attribute the comments to [Taylor] at all. From their responses to the questioning, it was evident that they had brushed off the comments and diligently followed the Court’s admonishments to avoid outside publicity and not consider *348anything they heard outside the courtroom. To the extent there might be some prejudice, the Court is convinced it can be cured by an instruction.

“Based upon its interviews with jurors,” the district court concluded, it was “confident that the verdict ... will not be based on emotion.” The district court further stated that Taylor’s request for a hearing to investigate potential juror prejudice had béen satisfied by the district court’s juror interviews, “especially given that defense counsel proposed the arrangement.” The court also reasoned that further questioning of the jurors “would obviously be detrimental to [Taylor]” because it “could bring the comments to the forefront of jurors’ minds and cause jurors to connect the statements to [Taylor].” Taylor appeals the district court’s decision, but only with respect to Juror 1.

The district court did not abuse its discretion by declining to ask Juror 1 explicitly about bias resulting from her exposure to Taylor’s “redneck” remark, although an explicit question certainly would have been preferable. When a juror is exposed to unauthorized communications involving a case, the district court must “determine the circumstances [of the communications], the impact [of the communications] upon the juror, and whether or not [the communications were] prejudicial.” Remmer v. United States, 347 U.S. 227, 229-30, 74 S.Ct. 450, 98 L.Ed. 654 (1954). The district court retains considerable discretion in deciding how to conduct such an inquiry, which our decisions have labeled a “Remmer hearing.” See Mays v. Chandler, 342 Fed.Appx. 159, 166-67 (6th Cir.2009). This discretion serves at least two important functions. First, it recognizes that district courts are uniquely qualified to ascertain the communication’s effect on the jurors and to determine the procedures that should be employed to deal with it. See, e.g., Evans v. Young, 854 F.2d 1081, 1084 (7th Cir.1988); United States v. Almonte, 594 F.2d 261, 266 (1st Cir.1979). Second, some degree of discretion is the only alternative to the sort of bright-line rule — for example, one requiring district courts to explicitly inquire whether a juror was prejudiced by a particular communication whenever that juror is found to have been exposed to the communication — that might actually exacerbate juror bias by “unnecessarily highlighting] the [communication] in the eyes of the jurors.” United States v. Mack, 729 F.3d 594, 606 (6th Cir.2013).

Upon learning that Taylor’s remark had been publicized, the district court asked every juror and every alternate about his or her exposure to Taylor’s remarks. The district court asked each juror and alternate who reported a direct familiarity with the remark whether he or she could remain impartial as a juror. Juror 1 had not heard about the remark directly; she knew about it only because some friends of hers had joked that they had not known she was a redneck. She drew certain inferences from those statements and speculated about their source, but did not actually know who made the remark. Juror l’s uncertainty about the nature of the remark and its origin left the district court with a difficult choice: leave the matter alone or press for more information and risk drawing further attention to the remark in Juror l’s mind. Particularly given that Juror 1 was uncertain about who had made the remark and what exactly had been said, the risk of exacerbating the comment’s potentially prejudicial effects may have outweighed the benefits of further questioning.

The district court was also best positioned to assess Juror l’s demeanor and behavior during the one-on-one interview, just as it was best positioned to assess the *349other jurors’ demeanor and behavior. When both parties subsequently asked the district court to inquire further of the jury, the court explicitly relied on its observations of the jurors’ demeanor and behavior to explain why further questioning was not only unnecessary but unwise. This does not appear to be a case in which the district court simply forgot to ask an important question. Indeed, the district court made a point of asking several other jurors whether they could be fair and impartial. The district court does not appear to have made a careless or thoughtless decision. To the contrary, despite multiple invitations to revisit the issue, the district court repeatedly reaffirmed its earlier conclusion that the remark had no meaningful effect on the jurors, including Juror 1, explaining that the jurors “who heard the remarks did not seem biased by them,” and that an alternate’s testimony suggested Taylor’s “comment sparked jokes and laughter among jurors; there was no indication jurors took the comment personally.” Finally, none of Juror l’s comments at the hearing inherently suggested any sort of bias arising from the allegedly prejudicial remark. Rather, she appeared to have brushed the remark off and to have been more concerned about other aspects of the case. That view is certainly consistent with the alternate juror’s observation that the remark had no meaningful effect on the jury’s members. Nothing in the record, in short, establishes the sort of obvious prejudice that would have required the district court to question Juror 1 further. Accordingly, Taylor has not carried his burden of establishing that the district court abused its discretion.

The cases that Taylor cites — particularly United States v. Davis, 177 F.3d 552 (6th Cir.1999) and United States v. Herndon, 156 F.3d 629 (6th Cir.1998)—are different. In Davis, for example, the court learned that one juror had found out from an outside source that the people in his part of town were aware of the juror’s jury service and were discussing the juror’s role in the proceedings. 177 F.3d at 556. The court also learned that the juror had shared the information — and his concerns for his safety — with other members of the jury. Id. Rather than investigating the extent to which the out-of-court communication and the juror’s unauthorized communications about the case had affected the remaining jurors, however, the district court simply discharged the troubled juror and went ahead with the case. Id. We vacated the Davis defendants’ convictions and remanded for a post-trial hearing at which the defendants would bear the burden of demonstrating the existence of jury taint, without which a retrial would not be necessary. Id. at 557. Our holding was based on the fact that the district court had made no inquiry into potential juror prejudice from the improper communications. Id. We relied on the fact that the juror in question “was clearly motivated by fear of retaliation from the defendants,” and the “fact that a number of jury members openly agreed that a person in [the juror’s] predicament should seek to be removed” from the jury. Id. We also relied on the total absence of a court effort to delve into the extent to which the juror’s comments affected the other jurors’ deliberations. Id. Neither of those factors is present in Taylor’s case. In particular, the district court, at the parties’ behest, conducted individualized conferences with the jurors to determine whether they had been prejudiced by the reported remark.

Herndon is likewise distinguishable. There, a juror recalled during the trial that he may have had prior business dealings with the defendant, but the district court declined to investigate — or to allow counsel to investigate — the potential for prejudice arising from those dealings. *350Herndon, 156 F.3d at 632-33. Concluding that the prior business dealings constituted an extraneous influence upon the juror, we vacated the defendant’s sentence and remanded for a Remmer hearing, on the ground that the district court had not provided any opportunity for the defendant to prove actual bias. The district court in Herndon rejected the request of Hern-don’s counsel to question the juror, and conducted no questioning of its own. Id. at 632. The only hearing was at the sentencing hearing, more than three months after conviction, at which the juror was of course not present, and the nature of which this court described as clearly “not an investigation” but permitting the defendant to “make a record” for the purposes of appeal. Id. at 632-33, 637. In Taylor’s case, by contrast, the district court asked all of the jurors about the extent to which they were aware of the reported remark. The court’s course of interviewing without the presence of counsel was with the consent of counsel. The record of the interviews gave Taylor the required opportunity to establish juror bias. Herndon, therefore, does not require vacatur or remand.

Taylor’s argument focuses not on the total lack of an opportunity to establish bias, but on the district court’s not having questioned the jurors as extensively as Taylor would have liked. He points to United States v. Corrado, 227 F.3d 528 (6th Cir.2000), and Goins v. McKeen, 605 F.2d 947 (6th Cir.1979), to support his point, but those cases are distinguishable because the district courts’ investigations of potential prejudice plainly fell short of what the law requires. In Corrado, for instance, we remanded for a hearing where, in response to an outsider’s attempt to bribe an unidentified juror, the district court merely “directed three broadly-worded questions to the jury as a group and instructed any jurors who [felt they could not remain impartial] to bring themselves to the court’s attention by writing a note.” 227 F.3d at 536. The district court’s “ ‘minimalist’ approach,” we held, “was an inadequate response to the serious and credible allegations of extraneous influences on the jury in th[at] case.” Id. at 535. The jury in Goins was similarly exposed to a prejudicial out-of-court communication — this time a newspaper article containing details of the defendant’s plea process, which suggested that the defendant had negotiated to plead guilty to a lesser charge. 605 F.2d at 954. This evidence was “strongly probative of guilt.” Id. at 953. Because the communication contained the type of “extremely prejudicial information” that “rendered the circumstances [surrounding a potential verdict] inherently prejudicial,” we held that the district court had to do more than merely inquire of the jury — en masse and in open court — whether any of the jurors had been exposed to the prejudicial communication and whether those who had could still be fair and impartial fact-finders. Id. at 954. Here, by contrast, the district court made a deliberate and concerted effort to investigate potential juror prejudice. Not only did the district court question the jurors in camera, one at a time, but it repeatedly asked follow-up questions when the need arose. The district court’s efforts here, then, were not at all like the cursory inquiries that were not adequate in Corvado and Goins.

Finally, United States v. Walker, 1 F.3d 423 (6th Cir.1993), is distinguishable because of the highly prejudicial communication at issue in that ease. In Walker, jurors were mistakenly shown portions of a transcript that had not been entered into evidence. Id. at 427. The defendants explicitly requested that the district court allow them to question the jurors about the effect of the transcript on their ability *351to be impartial, which request the district court denied altogether. Id. at 430. We remanded, holding that “denying the reasonable request to inquire into the jurors’ states of mind” deprived the defendants “of the opportunity to meet their burden of proving actual juror bias,” and that the defendants were thereby denied a fair trial. Id. at 431. Walker involved double exposure to selected testimony — a problem whose power to improperly influence “has long been recognized.” Id. at 430. The “unauthorized and uncontrolled ‘read back’ ” in Walker “created a substantial potential for undue emphasis and a special hazard that limited testimony may be taken out of context.” Id. In contrast, the information in Taylor’s case did not involve information that was even indirectly relat-. ed to guilt or innocence. • Moreover, as the district court observed, when questioned individually, none of the jurors seemed bothered by the remark. Indeed, many of them (including Juror 1) had not even read the- coverage of the remark. By the time sentencing actually began, almost three weeks had passed since the media reported the remark, so that there was “every reason to believe any remarks jurors heard w[ould] be out of their minds.” These circumstances are materially different from those in Walker. ■

There is a world of difference between failing to conduct a Remmer hearing and failing to ask a particular question at that Remmer hearing. The former is flatly unconstitutional. But we have never held that a defendant is entitled to resentencing simply because the district court did not incant magic words during a Remmer hearing. All that Remmer requires is that the defendant be given an opportunity to prove juror bias, see Walker, 1 F.3d at 431, not that he be given the opportunity to prove juror bias based on answers to predetermined questions. Neither Taylor nor the dissent points us to any case holding that a Remmer hearing ceases to function as such if the district court fails to ask any of the jurors a particular question, such as whether a communication affected their ability to be fair and impartial in discharging their duties.

There are good reasons that the law does not flatly require district courts to ask any juror exposed to an out-of-court communication whether the communication biased her. A district court’s role in conducting a Remmer hearing is to “determine the circumstances [of the allegedly prejudicial communication], the impact thereof upon the juror and whether or not [the communication] was prejudicial.” Remmer, 347 U.S. at 229-30, 74 S.Ct. 450. Making those determinations necessarily entails consideration of more than just the substance of a juror’s answer; it requires a court to evaluate the juror’s mannerisms, deportment, tone, delivery, and other elements. The law does not require district courts to disregard those elements in many cases — even when they clearly indicate to the district court that the juror was not biased by the out-of-court communication — and, instead, press forward with questioning in a way that might actually magnify the significance of the prejudicial communication in the juror’s mind.

While it probably would have been preferable for the distript court to have asked Juror 1 if her knowledge of Taylor’s remark affected her ability to be impartial, an arguably questionable decision is not the same thing as an abuse of discretion. The record in this case does not establish that the district court’s decision not to inquire further of the foreperson was an abuse of discretion, and Taylor’s first claim is accordingly without merit.

Ground II: Exclusion of Testimony

Taylor argues that the district court erred in excluding some of the testi*352mony by James Aiken and Dr. Mark Cunningham, ostensibly regarding Taylor’s “future dangerousness.” His argument fails for two reasons. First, the testimony at issue was not responsive to the Government’s arguments about Taylor’s future dangerousness, so that the district court did not abuse its discretion in refusing to admit the testimony as rebuttal evidence. Second, because the excluded testimony did not pertain to Taylor’s character, behavior, or the circumstances of his crime, it did not constitute relevant mitigating evidence.

The context for the excluded proffered testimony is as follows. The Government relied on evidence from both the guilt and sentencing phases of trial to support the non-statutory aggravating factor of future dangerousness. In the guilt phase, the Government presented testimony about Taylor’s attempt to escape from jail. Taylor, Marshall, and Matthews were being held in the Hamilton County, Tennessee jail on their federal charges. Taylor and Marshall became involved in an escape plan with three older inmates: Steven Sza-bo, J.R. Uhuru, and Thaddeus Reed. According to Marshall, Szabo, Uhuru, and Reed told Taylor about their plan to escape, and Taylor told Marshall. The plan was for Uhuru to hit one guard, for Reed and Szabo to restrain two guards, and for Taylor and Marshall to take the guards’ radios and keys. Uhuru acquired a length of pipe and flattened it to use to cut through the wire mesh in the cell window. The group also saved sheets and tied them together to use to climb down from their third-floor cells. Taylor arranged for his mother to have a vehicle waiting outside. In his guilty plea agreement, Szabo stated that Taylor was the leader of the conspiracy to escape. In his testimony, Szabo indicated that Uhuru gathered the items needed and punched the guard. During the escape attempt, Taylor helped one of the older inmates hold an officer down. Guards broke up the fight and ordered the inmates back to their cells.

In the sentencing phase, the prosecution introduced a letter that Taylor wrote to Heather Hamilton on September 12, 2008. In the letter, Taylor predicted his conviction would be overturned and wrote: “So many people want to do something to [Marshall]. Next time I go to trial, I bet he won’t testify against me. Trust me on that.” The letter also described Taylor’s methods of identity theft. After Taylor learned that the jail was monitoring his calls and letters, he used another inmate’s name to send a letter to Hamilton. Finally, according to FBI Special Agent Melia, Marshall’s grandmother or godmother told him that someone tried to break into her house, and that a relative in Taylor’s neighborhood had every window broken out of her home.

Taylor’s counsel presented James Aiken to testify in mitigation. He stated that he had worked in prisons since 1971. He was warden of a South Carolina prison, worked with the United States Bureau of Prisons (“BOP”) and prisons in Indiana, the Virgin Islands, and the supermax prison in Florence, Colorado, and has consulted across the United States. Aiken explained that prisons use internal classification systems to separate the most aggressive members of the prison population from the least aggressive. He told the jury that, in the federal system, officials consider the inmate’s age, build, criminal background, medical condition, gang membership, institutional behavior, sentence, and type of crime. Of these factors, the sentence and type of crime are the most important. Aiken predicted that Taylor would go to a high security setting. There would be a gun between Taylor and the public as long as he lives, and he would be accounted for continuously.

*353Aiken was asked about Taylor’s escape attempt. Aiken said that the county jail’s security did not compare to that of a high security federal facility. Aiken testified that BOP protocols, experience, and design would make the chance of an escape from a federal facility extremely remote. The district court cut off defense counsel’s question about how the BOP separates inmates who have conflicts, such as gang members, and inmates who have testified against other inmates or committed crimes together, and counsel agreed to move on to another subject. Aiken stated that Taylor was five feet, eight inches tall, weighed one hundred and forty to one hundred and fifty pounds, and looked extremely youthful. The district court sustained the prosecution’s objections to questions about what daily life in prison would be like for Taylor as far as security is concerned at BOP institutions. In doing so, the court stated that defense counsel had

gone much further than the parameters of this witness’s testimony. He was called to the stand to testify that the witness would be transferred from Hamilton County Jail to a federal prison facility someplace — no one knows what facility he will be going to ... and that whatever facility he goes to will have better security than the Hamilton County Jail.

The district court subsequently elaborated:

The witness can testify that, based upon his experience, that the defendant will be transferred from the Hamilton County Jail to a Bureau of Prisons institution — no one knows what institution that will be at this point; that is unknowable — and that whatever institution he goes to will have better security than exists at the Hamilton County Jail.

Defense counsel objected.

On cross-examination, Aiken testified that Taylor’s behavior while in the county jail would not cause him to be placed in the highest level of security in the BOP. He also testified that even high security federal facilities have escape attempts. On redirect, Aiken said that the BOP had the most secure facility in the world, which kept total control of every aspect of the inmates’ lives, but that Taylor would not be sent to it. If Taylor were sentenced to death, he would be sent to death row at a high security prison. Also, if Taylor were sentenced to life, he would be sent to a high security prison. Aiken explained that escape-prevention security measures at a high security BOP facility include monitoring devices, barriers, and firepower. Inmates are subject to strip searches and their communications are monitored. BOP officers take note of inmates’ demeanor, cell contents, telephone calls, and television habits. There are special housing units for those who violate the rules.

Clinical and forensic psychologist Dr. Mark Cunningham also testified on Taylor’s behalf. Dr. Cunningham has practiced, researched, and taught psychology since 1983 and written about death row inmates, violence in prisons, and capital sentencing. He has also testified over 200 times in state and federal court. Dr. Cunningham interviewed eleven members of Taylor’s family to prepare for his mitigation testimony. He acknowledged that Taylor had a choice regarding his crime, but said that people do not all get the same choices. Factors that stand between a person and bad outcomes include an intact family, acceptance, affirmation, stability, structure, and modeling of positive values. A family history of addiction, psychological disorder, child abuse, abandonment, or instability can lead to bad outcomes.

Dr. Cunningham referred to a 1995 Department of Justice (DOJ) study that *354sought to identify risk factors for criminal activity. Dr. Cunningham gave over sixty transcript pages of testimony applying these factors to Taylor, testimony that was not interrupted by objection or court limitation. Dr. Cunningham said that from birth to age 6, Taylor was exposed to criminal behavior, substance abuse, family management problems, family conflict, and inappropriate parental attitudes to crime and substance abuse. From age 6 to adolescence, Taylor was exposed to violence and crime, changed households, firearms, media portrayals of violence, poor family management and conflict, poor parental attitudes and involvement in crime and substance abuse, lack of commitment to school, and delinquent and violent peers. Taylor’s family tree includes several members with criminal records, including an aunt, uncle, cousin, and older brother. Family members also abused alcohol and drugs, and had children out of wedlock. Dr. Cunningham was especially concerned by the fact that Taylor’s mother had a history of antisocial or criminal behavior because that trait is more strongly inherited from the mother than the father. Taylor’s mother accepted Taylor’s father’s criminal activity and did not guide or discipline her children. She worked up to twenty hours a day. Taylor’s father committed burglaries at age 12, shot someone when he was 13 or 14, committed murder at age 17, committed armed robbery at 19, and was in prison for theft and robbery from age 21 to 29. Taylor’s mother helped his father escape from prison, and helped Taylor in his attempted escape from jail.

As a child, Taylor had prison visitations with his father up until he was 3. He had no contact with his father again until age 8. In Taylor’s middle childhood, his father was open about his criminality and involved his family. Taylor’s father’s criminal activity included robbery, identification theft, and fraud. He had stacks of cash in his home and handguns in his house and car. Taylor and his brother moved in with their father. Taylor’s brother began dealing drugs as a teenager. He accumulated thousands of dollars and gave drug money to Taylor. Dr. Cunningham concluded that Taylor had numerous risk factors that correlate with increased violence and crime.

Dr. Cunningham next discussed Taylor’s “wiring.” He noted that Taylor was 18 or 19 at the time of the offense, and that the brain continues to develop until the age of 25. Adolescents are prone to both reactive impulsivity and judgment impulsivity. Dr. Cunningham testified that there is a high mortality rate in adolescents because they put themselves in dangerous situations. Young adults’ reward centers go “on line” before their braking systems. Dr. Cunningham reviewed the findings of studies on adolescent and young men, which showed them to be much more likely to be violent' than older males. The district court sustained the Government’s objection when Dr. Cunningham was asked about security levels in the BOP.

The district court permitted Taylor’s counsel to proffer testimony by both Aiken and Dr. Cunningham that was not presented to the jury. Aiken said that prisons use classification systems to manage inmate populations by separating violent, predatory inmates from the normal, passive prison population. He testified that he reviewed Taylor’s jail records as of January 2008. He said that Taylor would be looked at as prey in prison and would have to be protected until he became seasoned. Aiken stated that, based on Taylor’s crime and sentence, incapacitation would be the driving force of his security classification. Aiken said that Taylor would start at a penitentiary and could be placed in protective custody to guard against sexual assault, extortion, gang activity, and contraband. *355Because Taylor would not be returning to the community, the goal would be to stabilize him. He would be constantly watched, evaluated, and subject to searches for contraband. The prison would control every aspect of his daily life. Aiken asserted that the BOP can safely control, manage, and contain Taylor for the remainder of his life without undue risk to staff, inmates, and the general community. Aiken was asked about Taylor’s attempt to escape from the Hamilton County Jail. Aiken said that the BOP has higher security standards that would make an escape attempt less likely. Aiken gave the probability of a successful escape from a BOP facility as nil. He testified that United States Penitentiaries are high security and have no public access. There is always á gun between the prisoner and the community. Within a penitentiary, an inmate can be locked up in a secure housing unit (SHU) for disciplinary reasons.

In Dr. Cunningham’s proffered testimony, he said that a capital inmate sentenced to life in federal prison would be sent to a “United States Penitentiary or higher” security level facility. Federal prisons are termed minimum, low, medium, and high security, and there are SHUs within prisons. The most secure facility is a supermax in Colorado. A high security institution has double fencing, gun towers, patrols, perimeter detection devices, and multiple barriers. Dr. Cunningham reviewed the BOP’s procedures for dealing with dangerous prisoners. He opined that the BOP can securely house and control inmates. When asked about future dangerousness and rates of violence, Dr. Cunningham testified that the BOP groups inmates by their convictions, personality characteristics, and other means. He cited statistics indicating that, from 2004 to 2005, 94% of prisoners did not assault anyone. There were three homicides in a population of 20,000 inmates. Dr. Cunningham explained that a jury would not know how often violence occurs in the BOP and how effective that system may be in containing violence. He indicated that much of the differences in violence levels system-wide is accounted for by differences in institutions rather than differences in the prisoners. Also, homicide rates are significantly higher outside of prison than inside. Dr. Cunningham also testified that murderers are disciplined less than other inmates and are less likely to be violent in prison. For federal capital offenders sentenced to life, their rates of violence are similar to those of other inmates. State prisoners sentenced to life without parole are half as likely to commit assault. Dr. Cunningham drew three conclusions from DOJ publications: violence in the community is not strongly or consistently associated with violence in prison; a prisoner’s current offense, prior convictions, and escape history are weakly associated with misconduct; and the severity of the offense does not predict a prisoner’s adjustment to prison.

The district court found that Dr. Cunningham’s proffered testimony was far afield from his area of expertise, not relevant, and based on hearsay. The court concluded that the evidence had nothing to do with Taylor and was misleading, confusing, and unreliable. In arguing in favor of admitting the evidence, Taylor’s counsel argued among other things that “[mjost jurors think we’re talking about Club Fed. And the only way they can learn the difference is for a man like Mr. Aiken and Dr. Cunningham to tell them the difference.”

Rather than relying on oral rulings from the bench, the district court explained the basis for its exclusion of parts of the testimony of Mr. Aiken and Dr. Cunningham in a written memorandum. The explanation is thoughtful, respectful of relevant prece*356dent, and compelling. It is best appreciated by extended excerpt:

General prison security is not a proper mitigating factor because it is unrelated to the defendant’s background, record, or character, or any other circumstance of the offense, 18 U.S.C. § 3592(a)(8), or any other mitigating factor required by the Constitution and the FDPA. In United States v. Johnson, the Seventh Circuit concluded a defendant “should not have been permitted to present to the jury, as he was, evidence of the existence of maximum-security federal prisons decked out with control units, in order to establish a mitigating factor.” 223 F.3d 665, 674-75 (7th Cir.2000). “The argument that life in prison without parole, especially if it is spent in the prison’s control unit and thus in an approximation to solitary confinement, sufficiently achieves the objectives aimed at by the death penalty to make the latter otiose is an argument addressed to legislatures, not to a jury.” Id. at 675. In addition, the government cites Johnson for the argument that Defendant’s argument is inherently illogical, “as it amounts to saying that because this defendant is so dangerous, he does not deserve to be sentenced to death, since his dangerousness will assure his secure confinement. And its illogic shows that it is really an argument against the death penalty, period, since if this defendant should be spared because he is unusually dangerous, surely less dangerous murderers should not be executed either, even though, because they are less dangerous, they are less likely to be confined securely.” Id. While Johnson disapproved of this evidence as a mitigating factor, it approved of it in a proper case as rebuttal evidence to counter future dangerousness, which the government has alleged as a non-statutory aggravating factor in this case, although not of the nature involved in Johnson. Id. at 674-75.
Defendant quotes from the Supreme Court’s decision in Skipper v. South Carolina, which held “evidence that the defendant would not pose a danger if spared (but incarcerated) must be considered potentially mitigating.” 476 U.S. 1, 5, 106 S.Ct. 1669, 90 L.Ed.2d 1 (1986). But the evidence in Skipper concerned the defendant’s good behavior in prison, not security information about the prison. The Supreme Court has not addressed the right of a capital defendant to present evidence of prison security considerations. Schmitt v. Kelly, 189 Fed.Appx. 257, 265 (4th Cir.2006) (concluding state court’s refusal to allow such evidence was not an unreasonable application of federal law). However, a federal district court has joined the Seventh Circuit in holding such testimony is admissible in rebuttal:
Because it will be obvious to the jury that the Government’s proof of this factor is intended to show that Wilson, rather than criminals in general, is a future danger, it will be equally obvious to the jury that [the expert’s] testimony, though based on statistical evidence, is intended to shed light on Wilson’s future dangerousness, not on that of anyone else. And because the Government will argue for the death penalty based in part on Wilson’s alleged future dangerousness, this court must give Wilson “fair opportunity to present argument as to the adequacy of the information to establish the existence” of that aggravating factor. 18 U.S.C. § 3593(c).
United States v. Wilson, 493 F.Supp.2d 491, 508 (E.D.N.Y.2007). The Wilson court, however, prohibited testimony about any specific institutions as unduly *357speculative. Id. The evidence would be especially important in rebuttal because the FDPA and Constitution both require Defendant have a fair opportunity to rebut the government’s evidence. 18 U.S.C. § 3593(c); Gardner v. Florida, 430 U.S. 349, 362, 97 S.Ct. 1197, 51 L.Ed.2d 393 (1977); Davis v. Coyle, 475 F.3d 761, 771 (6th Cir.2007) (holding state court improperly refused to admit evidence of defendant’s adaptation to prison because it was clearly relevant to future dangerousness); see also Johnson, supra, 223 F.3d at 674-75.
In this case, however, the proffered testimony clearly was not rebuttal evidence. The FDPA allows Defendant to “rebut any information received at the hearing.” 18 U.S.C. § 3593(c). Although “the hearing” appears to refer to the sentencing phase, it can only logically be read as referring to the entire trial since the government’s evidence at the guilt/innocence phase is part of the basis for the jury’s sentencing decision. Thus, Defendant has the right to rebut information and evidence received through the entire trial. This right to rebut relates to information or evidence, not allegations. If the government alleges something but has no evidence to support it, the jury will ignore it and there is nothing to rebut.
Accordingly, it is important to examine exactly what evidence or information the government presented and not be confused by its allegations or claims. At trial, the government’s evidence regarding future dangerousness, in the light most favorable to the government, showed that Defendant was involved in an attempt to escape from custody; wrote a letter which contained a statement that could constitute a threat against codefendant and witness Joey Marshall; could have been involved in vandalism and damage to Marshall’s grandmother’s home; and was able to send letters to third parties from the Hamilton County Jail in someone else’s name. The government rested in the sentencing phase of the case so there is no possibility of other evidence in its case in chief. The above was the extent of the government’s evidence of future dangerousness, which Defendant is entitled to rebut. This evidence, if believed by a rational juror, could support a conclusion that Defendant might be involved in or might influence others to commit harm to others, including prison staff and other inmates. Based on the evidence, no rational fact finder could conclude Defendant is likely to personally injure someone while in custody, whether a fellow inmate or a prison staff member. This is especially true given the jurors’ observation of Defendant’s slight build and physique.
Prison security conditions do not have any relevance to the government’s evidence. The defense has set up a chimera, which they seek to destroy by introducing rebuttal testimony. Had this been a case where there was evidence the defendant was, to use Mr. Aiken’s terminology, a “predator” with a propensity to personally harm others, then the defense would have a better argument. With a proper witness, such evidence might be admissible. However, this is not that case.
Appropriate rebuttal to the government’s evidence in this case would be information that it would be impossible for Defendant to ever communicate with anyone outside of prison for the rest of his life or that he would not be able to influence any other inmates while in prison. And, of course, with the proper witness, such as a psychologist, testimony would be proper either as mitigating evidence or rebuttal that Defendant did *358not possess the personal qualities to harm others either directly or indirectly and would be unlikely to ever develop those personal qualities.
At the conclusion of the live proffer of Aiken’s testimony, considering the proper scope of rebuttal evidence, the Court indicated Aiken could testify that after the trial Defendant will be transferred from the county jail where he is currently detained, and from which he allegedly tried to escape, to a Bureau of Prisons facility, where security would be different and greater. But the Court prohibited Aiken from testifying about any specific institution Defendant may be sent to or about general conditions of federal prisons. Until the Bureau of Prisons selects a facility for Defendant, it would be speculative to testify about where Defendant may be sent and what security measures may be in place at such facilities during the future times in which Defendant will be incarcerated. Although there are commonalities among the security conditions at all federal prisons, these are unrelated to any mitigating factor. Because the conditions in federal prisons are unrelated to the circumstances of this offense or to Defendant, and because they do not rebut evidence or information introduced at trial, the Court prohibited most testimony about them.
Even though the Court had limited Mr. Aiken’s testimony as described, after the government had completed its cross-examination, the defense on redirect examination of Mr. Aiken brought out a great deal of information regarding Bureau of Prisons policies, procedures, general prison conditions and typical life in prison. The government did not object to this testimony even though it exceeded the permissible scope of the witness’s testimony, so it was presented to the jury.
After the testimony of Mr. Aiken, the defense called Dr. Mark Cunningham as a witness. Dr. Cunningham is a clinical and forensic psychologist and was tendered to the Court as an expert. He gave lengthy and extensive testimony regarding Defendant’s family, family history, development, and other matters regarding Defendant. This testimony included his expert opinion regarding the effect of various influences in Defendant’s life on Defendant’s development and life choices.
Subsequently, however, Dr. Cunningham began testifying as a fact witness regarding the federal Bureau of Prisons, particular federal prison facilities, and Bureau of Prisons policies, regulations and procedures, prison classification, prison security, and crime conditions in prisons. Upon objection by the Government the Court sustained the objection. Thereafter, the Court sent the jury out of the courtroom and provided the defense the opportunity to have Dr. Cunningham present his entire testimony to the Court. During this session out of the presence of the jury, Dr. Cunningham testified at considerable length and utilized a slide show pertaining to the rate of violence among various segments of the prison population, risk analysis among federal prisoners in capital cases, rates and correlations of prison violence, and the capability of the federal Bureau of Prisons to prevent inmate violence. With the benefit of Dr. Cunningham’s entire testimony, the Court reiterated its decision to sustain the objection. The Court explained that the objection was sustained for two reasons: in large measure the testimony was not relevant to any issues involved in this case and Dr. Cunningham was not qualified to give some of the testimony. For these reasons Dr. Cunningham’s testimony *359was unreliable, would mislead the jury about the true issues of future dangerousness involved in this ease, and would impermissibly confuse the jury.
The major problem with Dr. Cunningham’s disallowed testimony is its irrelevance to the issues in this case. The essence of his testimony is that individuals imprisoned in federal prisons are not likely to commit crimes. This is indisputably true but it has nothing to do with this case. As explained above, the government has introduced no evidence that Defendant is likely to pose a danger of personally injuring another inmate in prison or personally attacking a prison staff member. So even if Dr. Cunningham had offered his opinion, which he did not, that Defendant was not likely to personally injure another inmate or a staff member, such opinion testimony would not rebut anything involved in this case. Permitting his generalized testimony about prison crime would simply confuse and mislead the jury and invite a decision on an impermissible basis. For the same reason the Court disallowed testimony from Mr. Aiken regarding prison security, Dr. Cunningham’s testimony about the capability of the Bureau of Prisons to secure inmates is not relevant to any issues in this case.
The second problem with Dr. Cunningham’s disallowed testimony was that much of it is not expert testimony but rather is fact testimony, so he is testifying not as an expert but as just another fact witness. As a fact witness he must have direct or first hand knowledge of the facts about which he testifies. It is permissible for a witness to testify both as a fact witness and as an expert. But the case law indicates the Court must make it clear to the jury the distinction between the two roles. Even though the rules of evidence are relaxed in the sentencing phase of the trial, the Court is still obligated to ensure that the information being presented is minimally reliable, can be tested, and is not confusing or misleading. From Dr. Cunningham’s testimony, the facts he related often came from second, third, or fourth hand sources. Those sources are not available for questioning in this trial.
The third problem with Dr. Cunningham’s disallowed testimony was that some of it amounted to instructing the jury on the law. His mention of Bureau of Prisons policies, regulations, and procedures were matters of law that are properly within the province of the Court and that witnesses are not allowed to discuss in trials. To the extent such information would be relevant in a case, it should be provided to the jury by the Court and not by witnesses.
Moreover, Dr. Cunningham’s testimony had nothing to do with the personal characteristics or record of Defendant but rather was about an extraneous body, i.e., the federal Bureau of Prisons. The capability or lack of capability of the Bureau of Prisons says nothing about the individual background, record, or character of Defendant or the circumstances of this case. Dr. Cunningham’s testimony may concern the background, record, and character of the Bureau of Prisons and the employees of that institution, but not Defendant. And of course it can be expected that the capability of the Bureau of Prisons will vary over time and will even vary within particular constituent parts of the organization. This focus on the Bureau of Prisons ignores that we are concerned with a living, breathing human being with thoughts, passions, instincts, goals, ambitions, motivations, and all the other intangibles that make us human beings. *360Regardless of how capable the Bureau of Prisons is, a highly motivated individual who desires to harm others poses a threat. Perhaps it is a preventable threat but it is still a threat.
Lastly, because Dr. Cunningham’s disallowed testimony could be given verbatim in any capital case in the country without changing a single word, it runs afoul of what the Supreme Court said should be the foundation of capital sentencing: an individualized inquiry. Jones v. United States, 527 U.S. 373, 381, 119 S.Ct. 2090, 144 L.Ed.2d 370 (1999). Testimony regarding hundreds or thousands of prisoners in groups, in many unidentified prisons, in circumstances that the jury could never know precisely, would run the risk of the jury losing sight of its obligation to focus on the individual before it, the defendant, his character or record, and the circumstances of the offenses. Individualized sentencing of a capital defendant is one of the most important decisions the jurors will ever make in their own lives- and testimony regarding generalities of prison invites the jury to make decisions based upon group characteristics and assumptions. This presentation, as the Seventh Circuit said in Johnson, could be directed to Congress to convince it there is no need for the death penalty, rather than to a jury.

Taylor, 583 F.Supp.2d at 936-43.

There is not any abuse of discretion in this analysis.

The excluded portion of Aiken’s testimony was not relevant rebuttal evidence on future dangerousness, because none of it rebutted any of the Government’s future dangerousness arguments. At no point in the proceedings did the Government seriously contend — or even intimate — that Taylor was personally dangerous. Indeed, any such argument would have been patently absurd given Taylor’s slight stature — something the jurors could not only see for themselves, but on which Aiken remarked at one point during his testimony. The Government made the point even more explicitly during its closing argument:

Now, why is Rejón Taylor dangerous, ladies and gentlemen? He’s not dangerous because he’s a big, powerful physical presence. He’s not intimidating. He looks like the librarian aides that my mother used to employ at the high school library. He’s not scary-looking. It’s just the opposite. He’s more dán-gerous because he’s not scary-looking, ladies and gentlemen. He’s more dangerous because you don’t see him coming. He’s dangerous because he’s smart. He’s dangerous because he’s controlling. He’s dangerous because he never stops manipulating. He’s dangerous because he never stops conspiring. He’s dangerous because he’s a chameleon. He’s dangerous because he has no remorse. He’s dangerous because he has no concern for others. He’s dangerous because prison walls cannot contain or deter his influence, ladies and gentlemen. He can recruit. He can influence. He can trick anyone.

That the Government neither suggested nor implied that Taylor would be directly dangerous to others while in prison — indeed, the Government did the very opposite — distinguishes this case from Supreme Court cases holding that, when a capital defendant’s danger to the community has been raised, but the only sentencing alternative to death available to the jury is life imprisonment without the possibility of parole, the jury must be informed of the parole ineligibility. Simmons v. South Carolina, 512 U.S. 154, 114 S.Ct. 2187, 129 L.Ed.2d 133 (1994), was a split decision in which the operative opinion was the con*361currence of Justice O’Connor. That opinion stressed that “the State [had] sought to show that petitioner [was] a vicious predator who would pose a continuing threat to the community.” Id. at 175-76, 114 S.Ct. 2187 (O’Connor, J., concurring) (emphasis added). Reversal was required because “[a]lthough the only alternative sentence to death under state law was life imprisonment without possibility of parole, petitioner was not allowed to argue to the jury that he would never be released from prison, and the trial judge’s instruction did not communicate this information to the jury.” Id. at 178, 114 S.Ct. 2187. Whereas in Simmons, “[t]he State raised the specter of [the defendant’s] future dangerousness generally” and “advance[ed] generalized arguments regarding” that specter, id. at 165, 171, 114 S.Ct. 2187, here the Government made specific arguments about the way in which Taylor might be dangerous in the future, arguments that went exclusively to the notion that Taylor could be vicariously dangerous if given a life sentence. In short, precluding evidence of ineligibility for parole in Simmons gutted the defendant’s ability to respond to the very arguments made by the prosecutors. Cf. Kelly v. South Carolina, 534 U.S. 246, 248, 252-56, 122 S.Ct. 726, 151 L.Ed.2d 670 (2002). Of course the present case does not involve ineligibility for parole. More generally, the inability to respond to arguments made by prosecutors regarding future danger is not present in anything like the way that it was in Simmons and Kelly.

Aiken’s testimony was similarly irrelevant for purposes of mitigation, since mitigating evidence must bear on the defendant’s particular character or circumstances. See, e.g., Zant v. Stephens, 462 U.S. 862, 879, 103 S.Ct. 2733, 77 L.Ed.2d 235 (1983); Eddings v. Oklahoma, 455 U.S. 104, 112, 102 S.Ct. 869, 71 L.Ed.2d 1 (1982); Jurek v. Texas, 428 U.S. 262, 276, 96 S.Ct. 2950, 49 L.Ed.2d 929 (1976); Gregg v. Georgia, 428 U.S. 153, 189 n. 38, 96 S.Ct. 2909, 49 L.Ed.2d 859 (1976). We have noted that mitigating evidence “includes any aspect of a defendant’s character or record and any of the circumstances of the offense that the defendant proffers as a basis for a sentence less than death.” Carter v. Bell, 218 F.3d 581, 594 (6th Cir.2000). This statement in Carter relied on language from the Supreme Court’s decision in Lockett v. Ohio, 438 U.S. 586, 604, 98 S.Ct. 2954, 57 L.Ed.2d 973 (1978), which in turn recognized “the traditional authority of a court to exclude, as irrelevant, evidence not bearing on the defendant’s character, prior record, or the circumstances of his offense.” Id. at n. 12, 98 S.Ct. 2954. As the court below explained, testimony about general prison conditions and the anticipated effectiveness of security protocols is insufficiently individualized to meet that standard.

Taylor cites the Eleventh Circuit’s decision in United States v. Troya, 733 F.3d 1125 (11th Cir.2013), to support his contention that the district court should have admitted Aiken’s testimony as mitigating evidence. Taylor’s reliance on Troya, however, is misplaced. In Troya, the Eleventh Circuit held that the trial court erroneously excluded mitigation evidence that “(1) Troya had received his GED, which statistically lowered the rate of risk of violence in prison; (2) Troya’s age of 25 lessened his risk of violence in prison; (3) Troya was likely to make a positive adjustment to prison based on his frequent familial visits; and (4) Troya had been safely managed, notwithstanding his escape attempt, during previous stints in incarceration.” Id. at 1137. All of that evidence was particular to Troya, inasmuch as it spoke to his “character or record and any of the circumstances of the offense that *362[Troya] proffer[ed] as a basis for a sentence less than death.” Carter, 218 F.3d at 594 (quoting Lockett, 438 U.S. at 604, 98 S.Ct. 2954). Taylor, by contrast, has not identified any information about his character, record, or the circumstances of his offense that the district court barred Aiken from presenting to the jury. Consequently, Troya does not undermine the conclusion that the district court did not abuse its discretion in refusing to admit parts of Aiken’s testimony as mitigating evidence. The Eleventh Circuit’s analysis in Troya was moreover not necessary to its judgment, because the court held that omitting the evidence was in any event harmless error. 733 F.3d at 1138.

Dr. Cunningham’s testimony about inmate classification procedures and prison security was also not proper mitigating evidence, because it consisted almost entirely of the sort of generalized facts that could be said to apply to every death-eligible offender. For instance, Dr. Cunningham was to testify that inmates on death row have a low recidivism rate and are statistically no more likely than other inmates to be dangerous in prison. That testimony is not specific to Taylor’s crime or character; the same could be said of any defendant on death row. Indeed, Dr. Cunningham had offered virtually the same testimony in other cases, without regard to the personal characteristics or conduct of the defendant. See, e.g., id. The generalized nature of his testimony makes it an argument against the death penalty itself, not an argument for sparing a particular defendant from the death penalty. Johnson, 223 F.3d at 674. Mitigating evidence must be sufficiently particularized to the defendant on whose behalf it is offered — to his character, his behavior, or the nature of his offense — so as to cause the defendant or his crime to seem less severe or harsh to the jury. Lockett, 438 U.S. at 604 n. 12, 98 S.Ct. 2954; Carter, 218 F.3d at 594. Mitigating evidence must, as the name suggests, mitigate the heinousness of the crime or the harshness of the criminal. Evidence about inmate classification procedures and prison security conditions does not fulfill that purpose. For that reason, the district court did not abuse its discretion in disallowing Dr. Cunningham’s testimony on those subjects.

Taylor argues that Jurek, 482 U.S. at 276, 107 S.Ct. 2360, required the district court to admit Dr. Cunningham’s testimony so that the jury would “have before it all possible relevant information” on mitigation. But the Supreme Court in Jurek included important limiting language. Ju-rek does not command district courts to admit as mitigating evidence any information that might conceivably help a defendant’s case. Rather, Jurek holds that “[w]hat is essential is that the jury have before it all possible relevant information about the individual defendant whose fate it must determine.” Id. (emphasis added). Dr. Cunningham’s testimony had nothing to do with Taylor; it was information about the prison population in general. It follows that nothing in Jurek required the district court to admit the excluded testimony. Indeed, Supreme Court precedents consistently emphasized the importance of evidence regarding the defendant in particular, as opposed to whole populations. See, e.g., Zant, 462 U.S. at 879, 103 S.Ct. 2733; Eddings, 455 U.S. at 112, 102 S.Ct. 869. Dr. Cunningham’s testimony does not fit that bill. This is not to say that statistical evidence like Dr. Cunningham’s can never be admissible. Quite the contrary, it can be relevant and admissible in many scenarios. For example, a district court might reasonably consider Dr. Cunningham’s summary of BOP studies as relevant evidence to rebut allegations that an individual might be directly dangerous in the future. That, however, was not an *363option in this case, because nothing in the record suggested that Taylor had any propensity to direct dangerousness. Thus, Dr. Cunningham’s testimony was irrelevant as rebuttal evidence and too generalized to be used as mitigating evidence. Accordingly, the district court did not abuse its discretion in preventing the jury from considering the evidence.

We do not say that the district court was required to exclude the evidence, or even that it was preferable to exclude it. It may be that in the exercise of abundance of caution, a court will permit testimony on behalf of the defendant that the court could properly exclude. The admission of such testimony may be well within the discretion of the district court. But the district court is not an automaton that can only come to one right answer on any evidentiary issue just because the case is a capital case. While reasonable judges may have exercised their discretion differently with respect to the admission of the Aiken or Cunningham testimony, the district court here did- not abuse its discretion.

Ground III: Hearsay at Sentencing

At sentencing, the Government called FBI Agent James Melia to testify about an interview he had conducted with Joey Marshall. According to Agent Melia, Marshall had reported that “about a week or so after the murder,” Taylor had said “that his heart couldn’t take it anymore, that he had to get out of town, and that he had been robbing drug dealers to try to raise money to get, and stay, out of town.” Later on in Agent Melia’s testimony, the Government asked him whether he had “any information about whether or not anything’s happened to Mr. Marshall’s family since he testified in this trial?” Agent Melia thereupon told the jury that Marshall’s grandmother had called Agent Melia to tell him that “it appeared that somebody had tried to break into her residence,” and that all of the windows in another of Marshall’s relative’s houses “had been broken out of the house.” Taylor contends that Agent Melia’s testimony on those points consisted of “vague, unconfirmed, multiple hearsay allegations,” and that the district court erred by admitting it.

The district court did not abuse its discretion in admitting Agent Melia’s testimony about his interviews with Marshall and Marshall’s grandmother. The Federal Rules' of Evidence do not apply at the sentencing phase of the trial, and accordingly the district court could properly admit hearsay evidence. See 18 U.S.C. § 3593(c); Sears v. Upton, 561 U.S. 945, 950, 130 S.Ct. 3259, 177 L.Ed.2d 1025 (2010). The only remaining question, then, is whether the district court could have reasonably found that Agent Melia’s testimony was relevant and reliable. See Sears, 561 U.S. at 950 n. 6, 130 S.Ct. 3259; Green v. Georgia, 442 U.S. 95, 97, 99 S.Ct. 2150, 60 L.Ed.2d 738 (1979). Agent Melia’s testimony about Taylor’s “my heart can’t take it” comment was relevant because it rebutted defense counsel’s suggestion, made during the trial’s guilt phase, that Taylor’s comment evidenced remorse. Agent Melia’s testimony was reliable enough because it was based on an official 302 Form that he had completed immediately after interviewing Marshall. That Marshall may have lied in some of his responses to Agent Melia’s questioning does not make Agent Melia’s testimony about his interview with Marshall or the contents of the Form unreliable. Whether Marshall’s statements to Agent Melia were truthful is a question that the jury could properly weigh for itself.

Agent Melia’s testimony about his conversations with Marshall’s grandmother was similarly admissible at sentencing. In the first place, Agent Melia’s testimony *364was seemingly corroborated by other evidence before the jury, including Taylor’s letter to his girlfriend. In that letter, Taylor not-so-subtly observed that “[s]o many people wants to do something to [Marshall]. Next time I go to trial, I bet he won’t testify against me. Trust me on that!” Taylor’s letter could fairly be read to hint at his familiarity with — and possible role in — reprisals against Marshall for testifying. Not only could the jury use that evidence to put Agent Melia’s testimony in context, but the jury could also use the evidence to assess the testimony’s reliability. Agent Melia’s testimony was relevant, moreover, to Taylor’s degree of remorse and to the degree to which Taylor was dangerous.

Finally, the district court could have reasonably concluded that the probative value of Agent Melia’s testimony was not outweighed by the risk that it would unfairly prejudice, confuse, or mislead the jury. When Agent Melia testified at sentencing, the jury had already sat through weeks of trial testimony and had convicted Taylor of kidnapping and murder. It had plenty of context, in other words, with which to evaluate Agent Melia’s testimony. Nothing Agent Melia said during his time on the witness stand — particularly given that his testimony pertained largely to minor offenses — was likely to confuse, mislead, or unfairly prejudice the jury. Accordingly, the district court did not err in admitting the challenged testimony.

Ground IV: Propriety of the Jury Instructions

There was also no reversible error in the jury instructions. Taylor argues that the district court “unfairly skewed its instructions towards the government, invaded the jury’s function, and undermined the jury’s consideration of mitigating factors.” In its instructions to the jury, the district court briefly summarized the evidence that the Government had introduced to support each proposed non-statutory aggravating factor. The district court refused, however, to enumerate in its instructions each of the thirty-three mitigating factors Taylor had proposed, let alone to summarize the evidence that Taylor had introduced in support of each factor. Taylor contends that, by summarizing the evidence supporting the Government’s position on aggravating factors and refusing even to list the factors that Taylor proffered in support of mitigation, the district court’s instructions prejudiced the jury toward a sentence of death.

Although the difference in the district court’s presentation of aggravating and mitigating factors was questionable, there was nothing confusing or misleading in the terms of the instructions. In the end, Taylor’s argument ultimately fails because he has not shown that the district court’s instructions actually prejudiced the jury. A jury instruction supports overturning a sentence only if the jury instructions, as a whole, were confusing, misleading, or — most pertinent to this case — prejudicial. Radvansky v. City of Olmsted Falls, 496 F.3d 609, 617 (6th Cir.2007). Nothing in the record indicates that the jury was actually prejudiced by the district court’s instructions in this case. To the contrary, the record indicates that the jury gave balanced, individualized consideration to each factor— aggravating and mitigating alike. With respect to the Government’s three proposed non-statutory aggravating factors, for example, the jury found that the Government had carried its burden with respect to one, but not with respect to two others, notwithstanding that the district court had, in instructing the jury, summarized the evidence supporting each factor. Obviously, then, the jury did not simply accept that every piece of evidence allud*365ed. to in the instructions was valid or sufficient to establish a given aggravating factor. Rather, the jury appears to have undertaken an independent analysis of the evidence, reaching its own varied conclusions. Under those circumstances, Taylor has not shown prejudice from the district court’s marshalling of evidence of non-statutory aggravating factors.

Likewise, Taylor has not shown that the jury was prejudiced by the district court’s refusal to list each of the thirty-three mitigating factors that Taylor proposed. In the first place, the district court’s instructions expressly referred the jury to the verdict form, which included a list of all of Taylor’s proposed mitigating factors. Jurors, we presume, follow the instructions that they are given, Penry v. Johnson, 532 U.S. 782, 799, 121 S.Ct. 1910, 150 L.Ed.2d 9 (2001), including, in this case, the instruction to consider carefully the evidence supporting each of the mitigating factors before voting on them. The jury’s actual votes on the individual mitigating factors further suggest that the jury gave thoughtful, particularized consideration to each of the factors that Taylor proposed. For example, all twelve jurors found in favor of sixteen of Taylor’s proposed mitigating factors, and all twelve jurors rejected twelve of Taylor’s other proposed mitigating factors. Differing numbers of jurors — nine, one, two, five, and three — found in favor of the remaining five proposed mitigating factors. The record therefore undermines Taylor’s suggestion that the district court’s failure to list the mitigating factors caused the jury to devalue those factors.

To be sure, a district court treads on thin ice when it summarizes evidence in instructing a jury, especially when the court does not treat the evidence similarly on both sides of a case. See United States v. Mundy, 539 F.3d 154, 158-59 (2d Cir.2008). That the practice is inadvisable, however, does not mean that it is invariably an abuse of discretion. Because Taylor has not carried his burden of showing prejudice from the district court’s descriptions of aggravating and mitigating factors, the district court’s instructions do not support overturning his sentence.

Ground V: Prosecutorial Misconduct During Closing Argument

Next, Taylor argues that his sentence must be reversed because, at summation, the prosecutor described Taylor in derogatory terms and showed the jury a gut-wrenching photo of Luck’s corpse at autopsy. None of the prosecutor’s conduct, however, warrants reversal.

During his summation, the prosecutor repeatedly compared Taylor to predatory figures. At one point, for example, the prosecutor said that Taylor “had been stalking this man and victimizing this man for over a year.” Later, the prosecutor described Taylor as “a remorseless and relentless hunter,” and told the jury that Taylor was “dangerous because he’s a chameleon.” The prosecutor then alluded to Robert Louis Stevenson’s classic short story, telling the jury that Taylor “is dangerous because he always looks like Dr. Jekyll but his mind, his mind, ladies and gentlemen, is always working like Mr. Hyde.” The prosecutor followed that comparison by advising the jury that it had “the obligation to be the sheep dog that protects the sheep, and even sometimes the wolves, from the wolf,” i.e., from Taylor.

The line between zealous prosecution and prosecutorial misconduct is sometimes a fíne one, but on balance the prosecutor did not overstep it here. A prosecutor may not make statements for the sole purpose of inflaming the jury’s passions against the defendant. Wogenstahl v. Mitchell, 668 F.3d 307, 333 (6th *366Cir.2012). Prosecutors are not categorically barred, however, from using rhetorical devices reasonably calculated to make relevant points, which is what the prosecutor did here. For instance, in emphasizing that Taylor had stalked and hunted Luck for an extended period of time, the prosecutor was not embellishing the record; he was merely stating a fact borne out by the Government’s evidence: that, in preparing for their crime, Taylor and his co-conspirators had taken the time to learn Luck’s schedule and whereabouts. That many people find the notion of hunting and stalking a human being uncomfortable does not make the prosecutor’s statements inaccurate. “Murder trials tend to involve grisly facts, and although prosecutors may not exploit the disquieting nature of those facts for verboten ends, prosecutors have no affirmative obligation to conceal or shy away from those facts.” Clarke v. Warren, 556 Fed.Appx. 396, 407-08 (6th Cir.2014).

The prosecutor also did not engage in misconduct by comparing Taylor to Mr. Hyde. Nothing in our precedents prohibits the use of literary allusions, particularly not when context makes clear that those allusions serve as near-universally-recognized shorthand for otherwise permissible commentary. There would be no issue, for example, had the prosecutor told the jury that Taylor was “superficially innocent-looking, but constantly scheming and manipulating.” Indeed, that very point — that Taylor, a seemingly harmless person, was actually a dangerous man — was related to the Government’s argument on future dangerousness and was thus relevant. The prosecutor could make that relevant point by referring to a well-known short story without committing prosecutorial misconduct.

Though close to the edge, it was also not misconduct for the prosecutor to analogize Taylor to a wolf or a chameleon. Prosecutors may usé the vernacular to express relevant thoughts. There is no reason, for example, that a prosecutor cannot use “chameleon” as shorthand to describe “a person given to often expedient or facile change in ideas or character.” See Merriam Webster’s Collegiate Dictionary 190 (10th ed. 1997). Moreover, as Taylor concedes, plain error must be shown with respect to the use of these analogies, as defense counsel did not object. The Supreme Court and this court have several times found no reversible misconduct where prosecutors compared defendants to animals or predators of some other kind. For instance, in Darden v. Wainwright, 477 U.S. 168, 181-83, 106 S.Ct. 2464, 91 L.Ed.2d 144 (1986), prosecutor statements referring to the defendant as an “animal” who “shouldn’t be out of his cell unless he has a leash on him” did not deprive the defendant of a fair trial. In United States v. Wettstain, 618 F.3d 577, 588-90 (6th Cir.2010), the prosecutor’s description of defendants as “monsters” at the center of a community-wide methamphetamine “epidemic” was improper but not flagrant. In Bedford v. Collins, 567 F.3d 225, 234 (6th Cir.2009), the prosecutor’s description of the defendant as a “demon” was unprofessional but did not deprive the defendant of a fair trial. See also United States v. Davis, 514 F.3d 596, 616 (6th Cir.2008) (prosecutor called defendant a “predator”); United States v. Ebron, 683 F.3d 105, 142-43 (5th Cir.2012) (prosecutor called defendant a “predator” and compared him to lions and tigers in the jungle stalking animals to kill); Shepard v. Lane, 818 F.2d 615, 621 (7th Cir.1987) (prosecutor called petitioner “a liar,” “a dog,” and “an animal”).

It was also not misconduct for the prosecutor to show the jury a grisly photograph of Luck’s corpse. Taylor does not contend that the photograph was doctored, inaccu*367rate, or irrelevant; his prosecutorial misconduct argument with respect to the photograph is, in essence, that the prosecutor could have shown the jury a less gory photograph of Luck’s corpse. But gore, grisliness, and gruesomeness are inherent in most murder cases — including this one — and prosecutors are not required to hide that fact from the jury. Clarke, 556 Fed.Appx. at 407-08. There is no requirement that prosecutors use the least distressing exhibits in a murder ease.

The photograph at issue was relevant in that it directly rebutted one of defense counsel’s closing arguments, namely the argument that Taylor was a better person than the nature of his crime suggested. Defense counsel having made that argument, the prosecutor was within bounds to direct the jury’s attention back to the nature of Taylor’s crime. The prosecutor did not present the photograph simply for its shock value or for the sake of inciting the jury to reach an emotional verdict. Even if improper, the photograph presentation, as in Wogenstahl, 668 F.3d at 333, was not flagrant. The prosecutor used the photograph to rebut one of Taylor’s arguments at sentencing. That was not prosecutorial misconduct sufficient to overturn the jury’s verdict in this case.

Ground VI: Consideration of Future Dangerousness as an Aggravating Factor

The district court also did not err, as Taylor contends, by refusing to strike the aggravating factor of future dangerousness in prison. The Supreme Court has held that the defendant’s future dangerousness is a consideration on which the State may rely in seeking the death penalty, see Simmons, 512 U.S. at 175, 114 S.Ct. 2187 (O’Connor, J., concurring) (citing California v. Ramos, 463 U.S. 992, 1002-03, 103 S.Ct. 3446, 77 L.Ed.2d 1171 (1983)), including death penalty cases where there is no possibility of parole. See id. at 163 (plurality opinion); see also Davis v. Coyle, 475 F.3d 761, 771 (6th Cir.2007). That is what happened in this case.

Notwithstanding the likelihood that Taylor, if not executed, will spend his whole life in prison, the consideration of future dangerousness has clear relevance in this case. As explained above, the Government’s future dangerousness argument had nothing to do with Taylor’s propensity for physical violence. Rather, the Government argued that Taylor was dangerous because of his ability to get others — inside and outside of prison — to act violently. The district court’s refusal to strike future dangerousness in prison as an aggravating factor thus does not warrant setting aside Taylor’s death sentence.

Ground VII: Sufficiency of the Evidence of Future Dangerousness

Moreover, contrary to Taylor’s next argument, the record contains sufficient evidence from which the jury could rationally have concluded that Taylor would pose a danger if incarcerated. At trial, for example, the jury learned that Taylor had participated in an attempted jailbreak — involving shanks and sections of piping — in which two guards were assaulted. The Government also introduced evidence from which jurors might reasonably have inferred that, while in jail, Taylor orchestrated multiple acts of vandalism against the family of a witness who had testified against Taylor. Finally, the jury repeatedly saw evidence and heard testimony — including Taylor’s own words— showing a lack of remorse on Taylor’s part. This included letters stating that Taylor did not mind being in prison, and the recording of a phone call during which Taylor laughed about Luck and Luck’s fiancée.

In weighing the sufficiency of the evidence of future dangerousness, we draw all *368reasonable inferences in favor of the jury’s finding, United States v. Newsom, 452 F.3d 593, 608 (6th Cir.2006), such that we will overturn the jury’s finding of future dangerousness only where no rational trier of fact could have arrived at the conclusion reached by the jury. See Lewis v. Jeffers, 497 U.S. 764, 781, 110 S.Ct. 3092, 111 L.Ed.2d 606 (1990). That is, as we have repeatedly noted, a “heavy burden,” United States v. Jefferson, 149 F.3d 444, 445 (6th Cir.1998), and one that Taylor has not carried here, given the evidentiary sources supporting the jury’s finding.

Ground VIII: Sufficiency of the Evidence of Planning and Premeditation

Similarly, there was sufficient evidence to support the jury’s finding that Taylor engaged in substantial planning and premeditation before murdering Luck — one of the Government’s aggravating factors. Taylor’s argument on this point is that the evidence shows only that he planned to rob or, at most, kidnap Luck, not that he planned to murder Luck. But the Government introduced several pieces of evidence that support a finding that Taylor planned to kill Luck, not just kidnap him. The jury learned, for instance, that Taylor arrived at Luck’s house the morning of the crime not just with loaded guns, but also without any sort of mask or other clothing that might conceal his identity — unlike previous burglaries. The jury could reasonably have interpreted the lack of masks as evidence that this was not just another robbery attempt, and that Taylor had no intent of letting Luck out of the van alive, since Luck could otherwise have easily identified Taylor to the police. While inside Luck’s house the morning of the crime, Taylor allegedly observed documentation indicating that Luck was helping to procure a warrant for Taylor’s arrest. Most significantly, Taylor drove Luck for two hours to a rural area of a different state, all the while not bothering to conceal his identity. The jury could have reasonably concluded that Taylor’s conduct suggested that he intended to kill Luck rather than leave him alive to testify against Taylor in the future.

In rejecting Taylor’s argument on premeditation, the district court thoughtfully reasoned as follows:

The evidence showed Defendant had been watching Luck and knew when Luck took money from his restaurant to the bank. Defendant picked that time to confront Luck. Luck was then forced into his van, and driven by Defendant from Atlanta, Georgia, to Collegedale, Tennessee, where Defendant fatally shot Luck.... The jury could conclude Defendant killed Luck after substantial planning and premeditation based on the fact that he learned Luck’s routine and killed Luck after taking him on a two-hour drive, during which time Defendant would have had substantial time to plan and think about the murder. Viewing the evidence in the light most favorable to the government, a rational jury could conclude Defendant had decided to kill Luck sometime before the murder and that the events of the day of the murder were just the culmination of a well thought-out plan.

That analysis is sound.

Taylor’s actions after the murder supplied further bases for concluding that he had intended the murder. In particular, the jury heard testimony that Taylor was calm and methodical after killing Luck, rather than panicked or surprised. The jury learned, for example, that, after the murder, Taylor promptly flipped the license plate on his vehicle, took precautions to minimize suspicion while returning to Atlanta, and then went out to eat. The jury could reasonably have interpreted Taylor’s conduct as evidence that he ex*369pected the murder and was not rattled or surprised by it.

Precedent does not require that the Government prove that Taylor planned to kill Luck in precisely the manner that he did. Thus, the fact that Luck’s-struggle may have prompted Taylor to shoot Luck sooner than he might have intended does not mean that Taylor did not engage in substantial planning and premeditation regarding Luck’s ultimate demise. The Government provided sufficient evidence of planning and premeditation in this case, and so Taylor’s sufficiency argument fails with respect to that aggravator.

Ground IX: Refusal to Admit Execution Impact Testimony

The district court also did not abuse its discretion in refusing to admit what defense counsel refers to as “execution impact” evidence — evidence about the impact Taylor’s execution might have on his family and friends. The district court, in a written memo, gave several cogent reasons for not permitting such evidence. First, the district court rejected an asserted equivalence between admissible victim impact testimony and so-called “execution impact” testimony:

[Wjhile the FDPA specifically provides for victim impact evidence, 18 U.S.C. § 3593, there is no parallel provision allowing for execution impact evidence. Asking the jury to sentence a defendant to life is not mitigating evidence. United States v. Mitchell, 502 F.3d 931, 991 (9th Cir.2007) (approving of district court not allowing witnesses to offer opinions on what jury’s verdict should be); United States v. Caro, 461 F.Supp.2d. 459, 465 (W.D.Va.2006) (“[A]n express plea for mercy to the jury from a defendant’s witness is not mitigating evidence that could aid the jury in their decision making.”); see also Stenson v. Lambert, 504 F.3d 873, 892 (9th Cir.2007) (noting there are no federal cases requiring the admission of execution impact testimony); Jackson v. Dretke, 450 F.3d 614, 617-18 (5th Cir.2006); but see Jackson, 450 F.3d at 620 (Dennis, J., dissenting).

The district court also noted other problems with such testimony.

First, the testimony would be speculative. While testimony about the victim’s death is based on past events that witnesses can clearly and accurately testify about, the execution of a person in the future is a speculative event. Testimony about a future event would require a witness to imagine the effects of a prospective occurrence, not relate facts that have occurred. Given the vagaries of human life and the plethora of internal and external events that impact our lives, both positively and negatively, to ask a person to testify with any degree of certainty about the impact on one’s life of an event that may happen many years in the future is simply to ask for conjecture.
Second, there is a matter of fairness. Victim impact witnesses are prohibited from opining on the proper sentence. Booth v. Maryland, 482 U.S. 496, 509, 107 S.Ct. 2529, 96 L.Ed.2d 440 (1987), overruled on other grounds, Payne v. Tennessee, 501 U.S. 808, 827, 111 S.Ct. 2597, 115 L.Ed.2d 720 (1991); Fautenberry v. Mitchell, 515 F.3d 614, 638 (6th Cir.2008). The reason for prohibiting such opinion testimony is that “any decision to impose the death sentence must be, and appear to be, based on reason rather than caprice or emotion. The admission of these emotionally charged opinions as to what conclusions the jury should draw from the evidence clearly is inconsistent with the reasoned decision-making we require in capital cases.” Booth, 482 U.S. at 509, 107 S.Ct. 2529. *370Emotionally charged opinions should not form the basis for a verdict in a capital case regardless of whether the opinions favor death or life.
In Payne, where the Supreme Court held that victim impact evidence was not constitutionally prohibited, the Court emphasized the fairness in considering both the life of the defendant and the life of the victim. 501 U.S. at 827, 111 S.Ct. 2597 (“Justice, though due to the accused, is due to the accuser also. The concept of fairness must not be strained till it is narrowed to a filament. We are to keep the balance true.”). The parallel between the defendant and victim supports prohibiting witnesses from opining on their desired sentence, regardless of their preference.
In a very general sense, since the victim of a murder in a capital case is not alive to testify at the trial, victim impact testimony serves the purpose of allowing the jury to get a feel and sense of the victim’s life. The defendant in a capital case will always be available to offer the jury relevant information regarding the defendant’s life and the impact of any sentence through mitigation evidence, relevant testimony from witnesses, or the defendant’s own testimony or statement. Permitting testimony as to the impact on third parties at Defendant’s execution would have upset this symmetry.

Taylor now argues that he had a right to introduce evidence about the “nature and strength of [his] loving relationships with his relatives,” on the ground that such evidence constituted “relevant mitigating evidence” under Skipper v. South Carolina, 476 U.S. 1, 4, 106 S.Ct. 1669, 90 L.Ed.2d 1 (1986). See also 18 U.S.C. §§ 3592(a)(8), 3593(c). He contends that testimony from his family would have constituted “relevant circumstantial evidence of his character, background, and value as a human being.” The district court, however, never prevented Taylor’s family members from testifying about Taylor’s “character, background, and value as a human being.” To the contrary, Taylor’s family and friends were allowed to testify at length about those subjects. The district court merely prevented Taylor’s family and friends from testifying about the impact that his execution would have on them. In so doing, the district court was on firm legal footing. As the Ninth Circuit has observed, there are no cases requiring the admission of “execution impact” testimony, even though some courts have allowed it. Stenson v. Lambert, 504 F.3d 873, 891 (9th Cir.2007). The fact that some courts have allowed execution impact testimony does not mean that the district court abused its discretion by finding such testimony irrelevant in this case. For one thing, there are good reasons to believe that such testimony would not properly constitute mitigating evidence. As the Fifth Circuit has noted, “[b]eeause such evidence does not reflect on [the defen-' dant’s] background or character or the circumstances of his crime, the Supreme Court has never included friend/family impact testimony among other categories of mitigating evidence that must be admitted during a capital trial.” United States v. Snarr, 704 F.3d 368, 401 (5th Cir.2013). Similarly, in United States v. Hager, the Fourth Circuit observed that:

[A]llowing a capital defendant to argue execution impact as a mitigator is improper .... The capital defendant is available to offer the jury all the relevant information as to his life, background, character, and the impact any sentence will have on him. To allow testimony of the impact on third parties, however, does nothing to inform the jury on any of these matters and upsets the balance set forth in [Payne v. Tennessee, *371501 U.S. 808, 111 S.Ct. 2597, 115 L.Ed.2d 720 (1991) ].

721 F.3d 167, 194-95 (4th Cir.2013).

Taylor argues that execution impact evidence should be allowed to the same extent as victim impact evidence, but these arguments are without merit for the reasons explained by the district court. See also Snarr, 704 F.3d at 402. The district court accordingly did not abuse its discretion by prohibiting execution impact testimony.

Ground X: Responses to Jury Questions

Taylor next argues that the district court’s instructions to the jury in response to its questions during deliberations were incomplete, misleading, and coercive. The district court gave the jury its sentencing instructions on October 14, 2008. Taylor requested that the district court instruct the jury that, if the jurors could not agree unanimously on a sentence, the district court would sentence Taylor to life in prison. The district court, however, declined to give that instruction. That determination was fully supported by the case law. See Jones v. United States, 527 U.S. 373, 380-81, 119 S.Ct. 2090, 144 L.Ed.2d 370 (1999). The instructions that the court gave said that the jury would decide Taylor’s sentence, that Taylor would be sentenced to death if the jury unanimously voted for that sentence, and that Taylor would be sentenced to life in prison without the possibility of release if the jury unanimously voted for that sentence. The district court also instructed the jurors that they should try to work out their differences and should not hesitate to change their minds if they became convinced that other jurors were right, but that they should not change their minds just because other jurors saw things differently or just to get the case over with.

The jury began deliberations on October 15, 2008, then took several days off. Deliberations resumed on October 21. At 11:15 a.m. that day, the jury sent a note to the district court asking: “If the jury cannot reach a unanimous decision, what then?” The court convened the parties and proposed the following instruction:

You have asked what happens if the jury cannot reach a unanimous verdict. This is a question that is often asked in trials taking a considerable amount of time or involving difficult issues. It is not uncommon for there to be disagreement or for jurors to find it difficult to reach a unanimous decision. However, we can only ask that the jury try its best to work out its differences.

Taylor objected and requested that the district court advise the jury that if its decision was not unanimous the district court would sentence Taylor to life in prison. The district, court overruled Taylor’s objection and sent its instruction to the jury. In doing so, the district court reasoned as follows:

[T]he questions [from the jury] are susceptible to multiple interpretations. One interpretation could be that they are deadlocked and they’re asking what to do. Another interpretation could be they’re just asking out of curiosity. Another interpretation is that they’re having some difficulty but they have not reached deadlock now.
One of the advantages to the instructions the Court gives is, it allows them to tell us specifically what the issue is. If they’re deadlocked — this is a very intelligent jury — they’ll tell us that. If they need something else, they will tell us that, also. What the instruct — what the communication does, in essence, is to repeat what the Court told the jury at the end of the instructions — do their *372best, work hard, and try to work out their differences.

The district court thus did not read the jury’s question as an expression of deadlock.

At 2:40 p.m., the jury sent out a second note asking whether the case would result in a mistrial and whether the guilty verdict would be “null and void” if the jury could not reach a unanimous decision. After once again convening and consulting with the parties, the district court responded with the following instruction:

If the jury is unable to reach a unanimous decision, then that result would be reported to the court, and the jury would be discharged. Whether or not the jury would be able to reach a unanimous decision in the sentencing phase of the case would have no effect on the jury verdict regarding Defendant’s guilt.

Taylor once more requested — and the district court once more rejected his request — that the district court instruct the jury that Taylor would be sentenced to life in prison if the jury failed to reach a unanimous decision. The jury returned its verdict about an hour after it sent its second note.

The instructions that the district court gave were accurate. Federal law allows for just two sentences for defendants convicted of kidnapping resulting in death: life in prison or the death penalty. 18 U.S.C. § 1201(a). If the jury unanimously votes for either sentence, the district court is to impose that sentence. Id. § 3594. If the jury is unable to reach a unanimous verdict, the district court is to impose the least severe sentence authorized by law— in this case, life in prison. Id. The district court’s initial instruction to the jury — read to the jury before deliberations commenced, and given to the jury, in paper form, at the outset of deliberations — accurately summarized the consequences of a unanimous verdict in either direction. The district court’s response to the jury’s second question, moreover, properly informed the jury of the consequence relevant to the jury of the jury’s failure to reach a unanimous verdict: discharge of the jury without effect on the jury’s guilty verdict.

The district court acted within its discretion when it did not tell the jury about the sentence that Taylor would receive in the event that the jury deadlocked. Jones holds squarely that the Constitution does not require the jury to be instructed on the consequences of failure to reach unanimity in a capital case. Jones, 527 U.S. at 383, 119 S.Ct. 2090. The cases that Taylor cites do not require a different result merely because the jury has asked. In Hodges v. Epps, 648 F.3d 283 (5th Cir.2011), and Morris v. Woodford, 273 F.3d 826 (9th Cir.2001), federal courts of appeals vacated death sentences imposed after lower courts erroneously instructed the juries that the defendants might receive a sentence of life in prison with the possibility of parole, when, in fact, the minimum sentence for which the defendants were eligible was life in prison without the possibility of parole. Hodges, 648 F.3d at 289; Morris, 273 F.3d at 837. Hodges and Morris thus involved affirmatively inaccurate statements of the law. Taylor, however, does not argue that the district court misstated the law in his case, only that the district court might have offered a fuller explanation. We are aware of no case that requires that a question from the jury be answered with information properly omitted from jury instructions.

It is, of course, possible for a district court to abuse its discretion in delivering jury instructions, even if those instructions are accurate, if the instructions are unduly coercive. There was nothing coercive in the answers provided by the district court, however. This was not a case where the *373jury stated that it was deadlocked, such that mere refusal to call a mistrial could be deemed coercive. It was accordingly within the district court’s discretion not to give a full Allen charge, but rather to respond only to the questions asked. The district court reasonably determined that if the jury were deadlocked, it would say so.

Even if the district court’s answers were considered analogous to an Allen charge, which is questionable in the context of this case, the operative question would be whether, “in its context and under all the circumstances,” the district court’s jury charge was “coercive.” Jenkins v. United States, 380 U.S. 445, 446, 85 S.Ct. 1059, 13 L.Ed.2d 957 (1965) (per curiam). In this case, the instructions that the district court actually gave were not, under the totality of the circumstances, coercive. In the first place, Taylor points to nothing in the instructions given that either singled out the minority of jurors or encouraged anyone in the jury to abandon sincerely held views of the case. That is one basis for distinguishing this case from United States v. Scott, 547 F.2d 334 (6th Cir.1977), where the district court took the unusual step of telling the jury that failure to reach a verdict would only exacerbate a backlog in cases on the court’s docket.

Furthermore, nothing in the jury’s instructions even suggested that the jury was required to reach an agreement. To the contrary, the district court’s second follow-up instruction clearly explored the possibility of the jury’s not reaching a consensus. The jurors’ right to continue disagreeing “may be implicit in the charge,” see Williams v. Parke, 741 F.2d 847, 850 (6th Cir.1984), and that is the ease here.

Accordingly, the district court’s responses to the jury questions did not amount to reversible error.

Ground XI: Sufficiency of the Evidence of Carjacking

Taylor next claims that the Government did not produce sufficient evidence to convict him of carjacking. To convict a defendant of carjacking in violation of 18 U.S.C. § 2119(3), the Government must show “that the defendant, (1) with intent to cause death or serious bodily harm, (2) took a motor vehicle, (3) that had been transported, shipped, or received in interstate or foreign commerce, (4) from the person or presence of another (5) by force and violence or intimidation.” United States v. Fekete, 535 F.3d 471, 476 (6th Cir.2008). Taking the evidence — particularly Matthews’s statements and Marshall’s testimony — in the light most favorable to the Government, the jury could reasonably have concluded that the Government had established each element beyond a reasonable doubt.

On appeal, Taylor does not contest that he took Luck’s vehicle, that Luck’s vehicle had been transported in interstate commerce, that he took Luck’s vehicle from Luck, or that he took Luck’s vehicle using intimidation. Thus, there is no question as to four of the five elements of carjacking under § 2119. Taylor’s only argument with respect to the jury’s carjacking verdict is that the Government did not prove that he took Luck’s car “with intent to cause death or serious bodily harm.” Matthews and Marshall, however, supplied ample evidence from which the jury could reasonably conclude otherwise. For example, Matthews stated that he, Marshall, and Taylor confronted Luck with weapons and deliberately forced Luck into his van; that Matthews trained his loaded gun on Luck as Taylor got into the van and drove it into Tennessee; and that Matthews and Taylor subsequently shot Luck. Marshall’s testimony, moreover, supplied a reason for Taylor to have acted *374with intent to cause death or serious bodily-harm, namely that Taylor had seen papers in Luck’s house suggesting that Luck had helped obtain an arrest warrant for Taylor. Based on Matthews’s and Marshall’s statements, the jury could have reasonably concluded that Taylor, brandishing the gun that he later used to kill Luck, intended to seriously harm or kill Luck when he and his co-conspirators forcibly took Luck’s van.

Taylor argues that carjacking requires the perpetrator to use or threaten force because the perpetrator wants to take possession of the vehicle, and that the Government never established that Taylor and his co-conspirators threatened Luck in order to take Luck’s van. Section 2119 contains no such requirement, however. To the contrary, § 2119 requires only that a defendant have acted “with intent to cause death or serious bodily harm.” We have never imported a “purpose” component into § 2119’s specific intent element. Instead, our precedents hold that § 2119’s specific intent element is satisfied “if a defendant brandishes a firearm and (1) physically touches the carjacking victim, or (2) there is direct proof that the firearm was loaded.” United States v. Washington, 714 F.3d 962, 968 (6th Cir.2013). Matthews had given a statement that he, Taylor, and Marshall “confronted Luck with firearms,” including a loaded 9mm semi-automatic pistol, to force Luck into the back of the van. Nothing more is required to find the requisite specific intent under § 2119.

Taylor points to a handful of cases ostensibly supporting his position that there was insufficient evidence to support the jury’s carjacking verdict, but those cases are distinguishable. In United States v. Applewhaite, 195 F.3d 679 (3d Cir.1999), the defendants assaulted the victim, threw his body into his van, and drove away in the van. There was no proof, however, that the defendants threatened or used force to take the victim’s van. Id. at 686. The victim was dead when the van was taken. Similarly, in United States v. Harris, 420 F.3d 467 (5th Cir.2005), it was uncontested that the defendant killed the victim and took the victim’s vehicle, but there was insufficient evidence to show that the defendant had the intent to kill or harm the victim at the moment that he took the vehicle. Id. at 473-75. Neither case, then, establishes any principle requiring overturning Taylor’s carjacking conviction.

Ground XII: Asserted Need for Nationwide Proportionality Review

Taylor next argues that, because juries have declined to impose capital punishment in cases comparable to Taylor’s, his sentence is constitutionally disproportionate, particularly where there is no statutory provision for a nationwide-proportionality review of federal capital sentences. No case is cited that supports such a holding. Indeed, the most relevant chses go the other way. The Supreme Court in Pulley v. Harris, 465 U.S. 37, 104 S.Ct. 871, 79 L.Ed.2d 29 (1984), held that California’s capital punishment scheme was not required to have statewide proportionality review; the possibility of state supreme court review was sufficient. The Court relied on its similar ruling in Jurek, 428 U.S. 262, 96 S.Ct. 2950. In an extensive analysis, moreover, our court has held en fianc that the defendant murderer-for-hire could be executed even if the mastermind of the murder was sentenced to life imprisonment by a different jury. Getsy v. Mitchell, 495 F.3d 295, 304-09 (6th Cir.2007). Taylor’s argument in this regard does not require reversal.

Ground XIII: Standard of Proof for Weighing Determination

Taylor argues that the reasonable doubt standard of proof should apply to the *375jury’s weighing determination. As Taylor acknowledges, however, that argument is foreclosed by our decision in United States v. Gabrion, 719 F.3d 511 (6th Cir.2013) (en banc), in which we held that the Federal Death Penalty Act does not require the jury to apply the reasonable doubt standard in weighing aggravating and mitigating factors. Id. at 531-33.

Ground XIV: Propriety of Venue Selection

Taylor also argues that the Government’s decision to prosecute him in the Eastern District of Tennessee was prompted by impermissible racial considerations. He suggests that the Government chose that venue because the jury pool in the Eastern District of Tennessee includes a higher percentage of Caucasian jurors than do the jury pools in the other venues where Taylor might have been tried. Taylor’s crime involved quintessential interstate activity, and the actual murder occurred in east Tennessee.

Taylor has submitted a prosecutor’s acknowledgment in a related case that the Government knew that it was not likely to secure a death sentence for Taylor in the state court system. Ordinarily, of course, there is nothing wrong with the Government’s selecting the venue in which the Government is most likely to secure what it believes is a just sentence. Taylor’s argument, however, is that juries in the Eastern District of Tennessee were more likely to sentence him to death because the jury pool in that district is made up largely of white would-be jurors. In other words, Taylor theorizes that a jury pool’s propensity to sentence black defendants to death is inextricably bound up with the racial composition of the jury pool, so that the Government’s selecting the most death-penalty-inclined venue in this case is an inherently race-conscious decision. That, Taylor contends, amounts to a Fifth Amendment equal protection violation.

To show that the Government’s venue decision violated his Fifth Amendment rights, Taylor must prove that the Government acted with discriminatory purpose and did so in a way that had a discriminatory effect. Wayte v. United States, 470 U.S. 598, 608, 105 S.Ct. 1524, 84 L.Ed.2d 547 (1985). He has supported neither element with anything other than speculation.

Taylor has produced no evidence that the Government made its charging decision based on the racial composition of the jury pool in any of the venues in which Taylor might have been tried. Prosecuto-rial decisions are presumptively legitimate, meaning that Taylor bears the burden of proving that the Government purposely discriminated here. At most, the evidence on which Taylor relies indicates that the Government knew that juries in the Southern Division of the Eastern District of Tennessee were more likely to return a death sentence than were jurors in some other jury pools. By itself, however, that fact does not establish that the Government brought the federal charge in east Tennessee because of the racial composition of the jury pool. Taylor’s contention in this regard does not warrant reversal.

Ground XV: Vagueness Challenge to 18 U.S.C. § 924(c)(3)(B)

Taylor’s final argument is that the Supreme Court’s decision in Johnson v. United States, — U.S. -, 135 S.Ct. 2551, 192 L.Ed.2d 569 (2015), compels the conclusion that 18 U.S.C. § 924(c)(3)(B), the statute supporting two of Taylor’s convictions, is unconstitutionally vague. Because § 924(c)(3)(B) is considerably narrower than the statute invalidated by the Court in Johnson, and because much of Johnson’s analysis does not apply to *376§ 924(c)(3)(B), Taylor’s argument in this regard is without merit.

Two of Taylor’s four convictions arose under § 924(j), which authorizes the death penalty where a person violates § 924(c) and causes the death of another person. Section 924(c) imposes mandatory minimum sentences for anyone who uses or carries a firearm during or in relation to a “crime of violence.” The statute defines crime of violence as:

[A]n offense that is a felony and—
(A) has an element the use, attempted use, or threatened use of physical force against the person or property of another, or
(B) that by its nature, involves a substantial risk that physical force against the person or property of another may be used in the course of committing the offense.

18 U.S.C. § 924(c)(3) (emphasis added). The jury found that Taylor murdered Luck in the course of committing two crimes of violence: kidnapping, in violation of 18 U.S.C. § 1201(a), and carjacking, in violation of 18 U.S.C. § 2119(3). At the time of trial, each crime qualified as a crime of violence as a matter of law under § 924(c)(3)(B), the italicized clause above.

Taylor argues that § 924(c)(3)(B) is not materially distinguishable from the residual clause of the Armed Career Criminal Act (ACCA), which the Supreme Court invalidated in Johnson as unconstitutionally vague. Johnson, 135 S.Ct. at 2563. The ACCA residual clause is part of the statute’s definition of “violent felony,” which includes:

[A]ny crime punishable by imprisonment for a term exceeding one year ... that—
(i) has as an element the use, attempted use, or threatened use of physical force against the person of another [the “use of physical force clause”]; or (ii) is burglary, arson, extortion, involves use of explosives [the “enumerated offense clause”], or otherwise involves conduct that presents a serious potential risk of physical injury to another [the “residual clause”].

18 U.S.C. § 924(e)(2)(B) (emphasis added).

Johnson does not require reversal of Taylor’s conviction, because several factors distinguish the ACCA residual clause from § 924(c)(3)(B). First, the statutory language of § 924(c)(3)(B) is distinctly narrower, especially in that it deals with physical force rather than physical injury. Second, the ACCA residual clause is linked to a confusing set of examples that plagued the Supreme Court in coming up with a coherent way to apply the clause, whereas there is no such weakness in § 924(c)(3)(B). Third, the Supreme Court reached its void-for-vagueness conclusion only after struggling mightily for nine years to come up with a coherent interpretation of the clause, whereas no such history has occurred with respect to § 924(c)(3)(B). Finally, the Supreme Court was clear in limiting its holding to the particular set of circumstances applying to the ACCA residual clause, and only some of those circumstances apply to § 924(c)(3)(B).

There are significant differences making the definition of “crime of violence” in § 924(c)(3)(B) narrower than the definition of “violent felony” in the ACCA residual clause. Whereas the ACCA residual clause merely requires conduct “that presents a serious potential risk of physical injury to another,” § 924(c)(3)(B) requires the risk “that physical force against the person or property of another may be used in the course of committing the offense.” Id. (emphasis added). Risk of physical force against a victim is much more defi*377nite than risk of physical injury to a victim. Further, by requiring that the risk of physical force arise “in the course of’ committing the offense, the language of § 924(e)(3)(B) effectively requires that the person who may potentially use physical force be the offender. Moreover, § 924(c)(3)(B) requires that the felony be one which “by its nature” involves the risk that the offender will use physical force. None of these narrowing aspects is present in the ACCA residual clause.

These are distinctions that made a difference in Johnson. The Johnson Court in part relied upon the wide judicial latitude permitted by the ACCA’s coverage of crimes that “involve[ ] conduct” presenting a serious risk of injury, language that did not limit a court’s inquiry to the elements of the crime. 135 S.Ct. at 2557. Section 924(c)(3)(B), by contrast, does not allow a court to consider risk-related conduct beyond that which is an element of the predicate crime since the provision covers offenses that “by [their] nature” involve a substantial risk that force may be used. The phrase “by its nature” indicates that a court’s analysis of whether there is a risk of force is confined to the offense itself. The Supreme Court has interpreted identical language in another statute, 18 U.S.C. § 16(b), in this way. See Leocal v. Ashcroft, 543 U.S. 1, 10, 125 S.Ct. 377, 160 L.Ed.2d 271 (2004).

Similarly, § 924(c)(3)(B) does not allow courts to consider conduct occurring after the crime has been committed. In Johnson, the Court explained that “the inclusion of burglary and extortion among the enumerated offenses preceding the residual clause confirms that” a court could evaluate the risk of injury arising after the crime has been completed, since “[t]he act of making an extortionate demand or breaking and entering into someone’s home does not, in and of itself, normally cause physical injury.” 135 S.Ct. at 2557. Section 924(c)(3)(B)’s requirement that physical force “be used in the course of committing the offense” permits no similar inquiry into conduct following the completion of the offense: under that statute, the force must be used and the risk must arise in order to effectuate the crime. Thus, unlike the ACCA residual clause, § 924(c)(3)(B) does not allow courts to consider “physical injury [that] is remote from the criminal act,” a consideration that supported the Court’s vagueness analysis in Johnson. Johnson, 135 S.Ct. at 2559.

Another independently compelling difference between the language in § 924(c)(3)(B) and the ACCA residual clause is the textual link in the latter clause by the word “otherwise” to four enumerated but diverse crimes. Johnson, 135 S.Ct. at 2558. The Johnson Court explained that by using the word “otherwise,” “the residual clause forces courts to interpret ‘serious potential risk’ in light of the four enumerated crimes — burglary, arson, extortion, and crimes involving the use of explosives.” Id. The Court further explained that gauging the level of risk required was difficult because the four listed crimes “are ‘far from clear in respect to the degree of risk each poses.’ ” Id. (quoting Begay v. United States, 553 U.S. 137, 143, 128 S.Ct. 1581, 170 L.Ed.2d 490 (2008)). Unlike the ACCA, § 924(c)(3)(B) does not complicate the level-of-risk inquiry by linking the “substantial risk” standard, through the word otherwise, “to a confusing list of examples.” Johnson, 135 S.Ct. at 2561. As a result, § 924(c)(3)(B) does not require analogizing the level of risk involved in a defendant’s conduct to burglary, arson, extortion, or the use of explosives.

A third consideration further militates against applying Johnson to this case. In Johnson, the Court recognized its “re*378peated attempts and repeated failures to craft a principled and objective standard out of the residual clause.” 135 S.Ct. at 2558. In the nine years preceding Johnson, the Court had applied four different analy-ses to the residual clause. See id. at 2558-59. These inconsistent decisions led to “pervasive disagreement [among lower federal courts] about the nature of the inquiry one is supposed to conduct and the kinds of factors one is supposed to consider.” Id. at 2560. By contrast, the Supreme Court has not unsuccessfully attempted on multiple occasions to articulate the standard applicable to the § 924(c)(3)(B) analysis. Nor is it correct, as Taylor argues, that the confusion about the ACCA in pre-Johnson Supreme Court decisions may be transferred to § 924(c)(3)(B), because much of the confusion in the ACCA cases concerned the four enumerated crimes that were linked to the residual clause. See Johnson, 135 S.Ct. at 2558-59.

Finally, the Johnson opinion itself stressed that its reasoning did not control other statutes that refer to predicate crimes. The Johnson majority addressed at some length the argument that the holding in that case would place in doubt “dozens of federal and state criminal laws[, like § 924(c)(3)(B), that] use terms like ‘substantial risk,’ ‘grave risk,’ and ‘unreasonable risk.’ ” Id. at 2561. The Court gave two reasons why that was not the case, and one of them directly distinguishes § 924(c)(3)(B):

Almost none of the cited laws links a phrase such as “substantial risk” to a confusing list of examples. “The phrase ‘shades of red,’ standing alone, does not generate confusion or unpredictability; but the phrase ‘fire-engine red, light pink, maroon, navy blue, or colors that otherwise involve shades of red’ assuredly does so.” James v. United States, 550 U.S. 192, 230, n. 7, 127 S.Ct. 1586, 167 L.Ed.2d 532 (2007) (Scalia, J., dissenting).

Johnson, 135 S.Ct. at 2561. Unlike the ACCA residual clause, however, § 924(c)(3)(B) does not include “a confusing list of examples.”

It is true that Johnson also relied in part on the fact that the ACCA residual clause, like § 924(c)(3)(B), requires the application of a categorical approach, which requires courts to look at the ordinary case of the predicate crime. Johnson, 135 S.Ct. at 2557-58. Taylor places primary emphasis on this similarity. But the Court explicitly refrained from flatly invalidating this type of analysis. Id. at 2561-62. In stating that the ordinary case analysis and the level-of-risk requirement “conspirefd] to make [the statute] unconstitutionally vague,” the Court ruled that the problem with the ACCA residual clause was that it combined an overbroad version of the categorical approach with other vague elements. Id. at 2557. The same is not true of § 924(c)(3)(B). In short, Johnson did not invalidate the ACCA residual clause because the clause employed an ordinary case analysis, but rather because of a greater sum of several uncertainties. Indeed, as the Court explained, although “the uncertainties in the [ACCA] residual clause may be tolerable in isolation, ... ‘their sum makes a task for us which at best could be only guesswork.’ ” Johnson, 135 S.Ct. at 2560 (quoting United States v. Evans, 333 U.S. 483, 495, 68 S.Ct. 634, 92 L.Ed. 823 (1948)).

Recent decisions, in addition, support distinguishing § 924(c)(3)(B) from the ACCA residual clause. In United States v. Amos, 501 F.3d 524, 527 (6th Cir.2007), we noted with respect to language in 18 U.S.C. § 16(b) identical to that in § 924(c)(3)(B) that “[t]he clause ‘used in the course of committing the offense,’ *379which does not appear in the ACCA, narrows the section 16(b) definition and distinguishes it from that in the ACCA.” See also United States v. Stout, 706 F.3d 704, 706 (6th Cir.2013). The Supreme Court and other circuits have also held that language identical to that used in § 924(c)(3)(B) is narrower than language identical to that used in the ACCA residual clause. See, e.g., Leocal, 543 U.S. at 10 n. 7, 125 S.Ct. 377 (comparing 18 U.S.C. § 16(b) with U.S.S.G. § 4B1.2(a)(2), which contains the same language as the ACCA residual clause); United States v. Serafin, 562 F.3d 1105, 1109, 1114 (10th Cir.2009) (comparing § 924(c)(3)(B) with U.S.S.G. § 4B1.2(a)(2) and with the ACCA residual clause).

We recognize that the Seventh and Ninth Circuits recently invalidated 18 U.S.C. § 16(b) under Johnson’s reasoning. See United States v. Vivas-Ceja, 808 F.3d 719, 723 (7th Cir.2015); Dimaya v. Lynch, 803 F.3d 1110, 1120 (9th Cir.2015). Although § 16(b) appears identical to § 924(c)(3)(B) in all material respects, neither decision changes our conclusion, for the reasons given above. Taylor’s argument that Johnson effectively invalidated § 924(c)(3)(B) is accordingly without merit.

The judgment of the district court is affirmed.

. In fact, Taylor had not used both terms together; the prosecution conflated two separate remarks from the same conversation.






Concurrence in Part

HELENE N. WHITE, Circuit Judge,

concurring in part and dissenting in part.

A death sentence is “unique [among criminal sentences] in its ... irrevocability,” Gregg v. Georgia, 428 U.S. 153, 187, 96 S.Ct. 2909, 49 L.Ed.2d 859 (1976) (plurality opinion), which makes it, of course, the “most severe punishment,” Roper v. Simmons, 543 U.S. 551, 568, 125 S.Ct. 1183, 161 L.Ed.2d 1 (2005). For this reason, the Eighth Amendment’s prohibition on “cruel and unusual punishments” and its protection of human dignity create a “heightened need for reliability in the determination that death is the appropriate punishment in a specific case.” Caldwell v. Mississippi, 472 U.S. 320, 323, 105 S.Ct. 2633, 86 L.Ed.2d 231 (1985) (internal quotation marks omitted).

Capital punishment “must be limited to those offenders who commit ‘a narrow category of the most serious crimes’ and whose extreme culpability makes them ‘the most deserving of execution,’ ” Roper, 543 U.S. at 568, 125 S.Ct. 1183 (quoting Atkins v. Virginia, 536 U.S. 304, 319, 122 S.Ct. 2242, 153 L.Ed.2d 335 (2002)); see also Kennedy v. Louisiana, 554 U.S. 407, 435, 439, 128 S.Ct. 2641, 171 L.Ed.2d 525 (2008) (holding that human dignity presumes both respect for the individual and a necessity to constrain capital punishment’s use). Thus, a death sentence is invalid and unconstitutional if a defendant’s crimes “cannot be said to have reflected a consciousness materially more depraved than that of any person guilty of murder.” (Godfrey v. Georgia, 446 U.S. 420, 433, 100 S.Ct. 1759, 64 L.Ed.2d 398 (1980) (internal quotation marks omitted)); see also Atkins, 536 U.S. at 319, 122 S.Ct. 2242 (holding that “the culpability of the average murderer is insufficient to justify the most extreme sanction available”).

The “underlying principle that the death penalty is reserved for a narrow category of crimes and offenders” is implemented throughout the capital sentencing process. Roper, 543 U.S. at 568-69, 125 S.Ct. 1183. Sentencing rules thus limit the death penalty to murderers most deserving of execution (in cases of crimes against individual persons), and protect a criminal defendant’s rights, including his right to a fair trial. See Eddings v. Oklahoma, 455 U.S. 104, 110-12, 102 S.Ct. 869, 71 L.Ed.2d 1 (1982). The Federal Death Penalty Act (FDPA) accordingly requires this court to “review the entire record in the case,” and *380directs us to examine the issues raised by the capital defendant, as well as consider “whether the sentence of death was imposed under the influence of passion, prejudice, or any other arbitrary factor and whether the evidence supports the special finding of the existence of an aggravating factor required to be considered.” 18 U.S.C. § 3595(c). A careful review of the record, measured against the Constitution’s guarantees and the FDPA’s directives, mandates that Taylor’s death sentence be vacated. Further, the Supreme Court’s decision in Johnson v. United States, - U.S. -, 135 S.Ct. 2551, 192 L.Ed.2d 569 (2015), compels the conclusion that 18 U.S.C. § 924(c)(3)(B) — which provides the definition of “crime of violence” underlying Taylor’s two § 924(j) convictions — is unconstitutionally vague, and that those convictions must be vacated. For these reasons, I respectfully dissent from Parts I, II, III, VIII, and XV of the majority opinion, join in Parts TV, V, IX, X, XI, XIII, and XIV, and concur separately in Part V.1

Ground I

The Sixth Amendment guarantees to criminal defendants the right to a fair trial by an impartial jury. U.S. Const. amend. VI; Irvin v. Dowd, 366 U.S. 717, 722, 81 S.Ct. 1639, 6 L.Ed.2d 751 (1961). Under Remmer v. United States, 347 U.S. 227, 74 S.Ct. 450, 98 L.Ed. 654 (1954), if a district court learns that a juror has been exposed to an unauthorized communication that may affect the juror’s partiality and thereby prejudice the defendant, the court must “determine the circumstances [of the unauthorized communication], the impact [of the communication] on the juror, and whether or not [the communication] was prejudicial, in a hearing with all interested parties permitted to participate.” Id. at 229-30, 74 S.Ct. 450 (emphasis added); cf. Maj. Op. 347-48 (omitting italicized language). Remmer thus imposes on a district court two distinct obligations: a duty to investigate allegations of a juror’s exposure to an unauthorized communication and a duty to determine whether the exposure violated the defendant’s constitutional rights. See United States v. Corrado, 227 F.3d 528, 535-36 (6th Cir.2000) (stating that “ ‘when there is a credible allegation of extraneous influences, the court must investigate sufficiently to assure itself that constitutional rights of the criminal defendant have not been violated’” (quoting United States v. Rigsby, 45 F.3d 120, 124-25 (6th Cir.1995))); United States v. Shackelford, 777 F.2d 1141, 1145 (6th Cir.1985) (holding that where “possible juror misconduct is brought to the trial judge’s attention he [or she] has a duty to investigate and to determine whether there may have been a violation of the [S]ixth [A]mendment”); see also Williams v. Bagley, 380 F.3d 932, 945 (6th Cir.2004) (observing that “[c]learly established Supreme Court precedent dictates that when a trial court is presented with evidence that an extrinsic influence has reached the jury which has a reasonable potential for tainting that jury, due process requires that the trial court take steps to determine what the effect of such extraneous information actually was on that jury” (internal quotation marks omitted)). “Where a col-orable claim of extraneous influence has been raised, ... a Remmer hearing is necessary to provide the defendant with the opportunity to prove actual bias.” United States v. Herndon, 156 F.3d 629, 635 (6th Cir.1998) (emphasis added) (inter*381nal quotation marks omitted); id. at 637 (reiterating that “the district court has a duty to investigate [a claim of alleged juror influence], and if the allegation is of an extraneous influence, then the district court should have conducted a Remmer hearing”); United States v. Davis, 177 F.3d 552, 557 (6th Cir.1999); see also Smith v. Phillips, 455 U.S. 209, 215, 102 S.Ct. 940, 71 L.Ed.2d 78 (1982) (observing that “the remedy for allegations of juror partiality is a hearing in which the defendant has the opportunity to prove actual bias” and discussing Remmer); United States v. Zelinka, 862 F.2d 92, 95-96 (6th Cir.1988) (holding that “when a defendant alleges that an unauthorized contact with a juror has tainted a trial, a hearing must be held”); Rigsby, 45 F.3d at 123 (stating that “we require a Remmer hearing in all cases involving an unauthorized communication with a juror or the jury from an outside source that presents a likelihood of affecting the verdict”); United States v. Pennell, 737 F.2d 521, 532 (6th Cir.1984) (observing that “Remmer ... controls the question of how the district court should proceed where ... allegations [of potential jury partiality] are made, i.e., a hearing must be held during which the defendant is entitled to be heard”).

Here, after individually interviewing each juror and alternate juror, the district court learned that most had indeed heard the reported “racist redneck” comment. The court thus should have provided Taylor a meaningful opportunity to prove actual bias. Because the court denied Taylor’s request for a Remmer hearing, it abused its discretion.2 See Herndon, 156 F.3d at 637 (holding that although “this court affords broad discretion to the district court in determining the type of investigation necessary to determine juror bias, the district court must provide the defendant a meaningful opportunity to prove the same”); see also Turner v. Murray, 476 U.S. 28, 36-37, 106 S.Ct. 1683, 90 L.Ed.2d 27 (1986) (plurality opinion) (“By refusing to question prospective jurors on racial prejudice, the trial judge failed to adequately protect petitioner’s constitutional right to an impartial jury.”).

The majority cites United States v. Mack, 729 F.3d 594, 606 (6th Cir.2013), to support its position that district courts are given discretion — as an alternative to a bright-line rule — because “requiring district courts to explicitly inquire whether a juror was prejudiced ... might actually exacerbate juror bias by ‘unnecessarily highlighting] the [communication] in the eyes of the jurors.’ ” Maj. Op. 348 (quoting Mack, 729 F.3d at 606). Contrary to the majority’s understanding of Mack, we did not there expressly reject a bright-line rule or even approve the district court’s decision not “to make an issue of it” so as not to “unnecessarily highlight the matter in the eyes of the jurors.” Mack, 729 F.3d at 606. Rather, the court noted the defendant never requested a Remmer hearing and concluded there was no plain error, explaining:

A defendant is entitled to a Remmer hearing to establish actual bias of a juror “only when the alleged contact presents a likelihood of affecting the verdict.” United States v. Frost, 125 F.3d 346, 377 (6th Cir.1997). “[N]o presumption of prejudice arises merely from the fact that improper contact occurred.” United States v. Davis, 177 F.3d 552, 557 (6th Cir.1999). Here, the district court discharged its duty to investigate and decide whether there may have been a violation of the Sixth Amendment right to a fair and impartial jury. See id. “[A] defendant who waits until appeal *382to request a hearing bears a heavy burden, since the defendant has thereby effectively deprived this court of any basis for concluding that a hearing would be necessary, and asks us to presume that the district court would not have acceded to such a request, and would have done so for erroneous reasons.” United States v. Walker, 160 F.3d 1078, 1083 (6th Cir.1998). Under the circumstances, no plain error exists that would justify reversal on this ground.

Id.

Unlike Mack, Taylor expressly requested a Remmer hearing, and the district court itself understood that the redneck remark was potentially prejudicial. Because the potentially prejudicial extrinsic communication presented a likelihood of affecting the verdict, this court’s precedents required the district court to afford Taylor a meaningful opportunity to prove actual bias. Moreover, the majority’s concern that further questioning could have exacerbated the comment’s potentially prejudicial effect3 fails to consider that two weeks after the juror interviews the district court itself reminded the jurors of the comment.4 Further, if a trial court can simply refuse to hold a Remmer hearing to avoid highlighting a prejudicial extraneous influence, the right to such a hearing is virtually nonexistent.

The majority also distinguishes United States v. Davis, 177 F.3d 552 (6th Cir.1999), and United States v. Herndon, 156 F.3d 629 (6th Cir.1998). Noticeably absent from the majority’s discussion of Davis is the Davis court’s reliance on a third factor: “the fact that the information that prompted the fear was provided by an extraneous source.” 177 F.3d at 557. This fact is critical because, as Davis recognized, although “a new trial will not be necessary every time a question of juror partiality is raised[, wjhere a colorable claim of extraneous influence has been raised, however, a ‘Remmer hearing’ must be held to afford the defendant an opportunity to establish actual bias.” 177 F.3d at 557 (internal quotation marks omitted). Such a claim is present here. Nevertheless, the majority concludes that the absence of the two other factors — that the juror was “clearly motivated by fear of retaliation from the defendants” and that “a number of jury members openly agreed that a person in [the juror’s] predicament should seek to be removed from the panel,” id. — adequately distinguishes Davis. But these were simply the circumstances of Davis; each case will necessarily involve different facts. It does not follow that the general rule that a Remmer hearing is required where a colorable claim of extraneous influence has been raised is not applicable. Moreover, it is not true that Davis relied on “the total absence” of an investigation as a basis for the court’s holding that the district court erred in not questioning the remaining jurors. Rather, the total lack of an investigation was the basis of the court’s holding that neither the trial court nor this court could determine whether the extraneous communications were harmless. 177 F.3d at 557. Finally, the majority’s statement that the district *383court here “conducted individualized conferences with the jurors specifically to determine whether they had been prejudiced by the reported remark” misapprehends the court’s and Taylor’s role. The court has a duty to investigate the allegation and Taylor has the burden to prove actual bias. By refusing to further question Juror One and denying Taylor an opportunity to do so, the court deprived Taylor a meaningful opportunity to prove bias.5

The majority distinguishes Herndon on the basis that here “the district court asked all of the jurors about the extent to which they were aware of the reported remark.” Maj. Op. 350. This characterization of the facts omits the critical fact that the court did not question Juror One (and only Juror One) whether she could continue to serve as an impartial juror. And Herndon suggests it is error to fail to obtain a juror’s personal assurance of impartiality when the juror is exposed to an unauthorized outside communication. 156 F.3d at 637 (observing that the trial “judge not only failed to adequately investigate the allegation of juror partiality, but the juror never personally assured the court of his impartiality”).

The majority concludes the transcripts of the court’s in camera interviews with the jurors “gave Taylor the required opportunity to establish juror bias.” Maj. Op. 350. Had the court questioned Juror One regarding her ability to remain impartial, I might agree. In any event, our precedents require a meaningful opportunity to prove actual bias in an evidentiary hearing. See Corrado, 227 F.3d at 536 (holding that “district court abused its discretion by failing to conduct an adequate evidentiary hearing into the allegations of extraneous influences on the jury pursuant to the holding in Remmer ” and remanding for a hearing at which the defendant “should be accorded the opportunity to question the jurors individually and under Qath”); Herndon, 156 F.3d at 637 (holding that “the district court must provide the defendant a meaningful opportunity to prove [juror bias]”); United States v. Walker, 1 F.3d 423, 431 (6th Cir.1993) (“By denying the reasonable request to inquire into the jurors’ states of mind, the defendants were deprived of the opportunity to meet their burden of proving actual juror bias----”); see also Remmer, 347 U.S. at 230, 74 S.Ct. 450 (holding that the district court “should determine the circumstances, the impact thereof upon the juror, and whether or not it was prejudicial, in a ■hearing with all interested parties permitted to participate”).

The majority correctly obseryes that Taylor focuses his appellate argument on the district court’s failure to further question Juror One. In other words, Taylor contends the district court failed to discharge its duty to investigate the impact of the outside communication on the juror’s ability to continue to serve as a fair and impartial juror. The import of asking the juror whether she could put the comment aside and decide the case based solely on the evidence and law given by the court is not inconsequential. A district court “may rely [on a juror’s assurance of continued impartiality] in deciding whether a defendant has satisfied the burden of proving actual prejudice,” if the court finds the juror’s assurance credible. United States v. Pennell, 737 F.2d 521, 533 (6th Cir.1984) (discussing Smith v. Phillips, 455 U.S. 209, 102 S.Ct. 940, 71 L.Ed.2d 78 (1982)). But *384absent such assurance, the court cannot fully assess whether a juror who is exposed to an outside communication can remain impartial. See Corrado, 227 F.3d at 537; United States v. Davis, 177 F.3d 552, 557 (6th Cir.1999) (holding that because “the [trial] judge did not inquire of the remaining jurors whether they were influenced in any manner, ... further inquiry seems not only appropriate, but necessary to ensure the impartiality of the jury”).6 For example, in Pennell, this court held the district court did not abuse its discretion in denying the defendant’s motion for a mistrial based on several jurors’ testimony that they had received late-night telephone calls threatening that they “had better find [the defendant] guilty,” because, unlike here, “the court immediately conducted a Remmer hearing ... and thoroughly questioned the contacted jurors on an individual basis and concluded that their assertions of unimpaired impartiality were worthy of belief.” Id. at 529, 534 (emphasis added).

In United States v. Walker, 1 F.3d 423 (6th Cir.1993), the district court inadvertently allowed inadmissible portions of deposition transcripts to enter the jury room. Id. at 427. After learning of the mistake, the district court interviewed the jurors individually in chambers and with counsel present to determine how the jury used the transcripts. Id. at 430. The court learned that the jurors did not read the inadmissible portions, but did read portions of testimony the court allowed the Government to show by video (but did not admit as an exhibit) and thus the jury was doubly exposed to the testimony. Id. at 426, 430. The court did not, however, ask the jurors whether the transcripts had influenced their ability to be fair. Id. at 430. Defendants suggested the court inquire of the jurors whether their double exposure would make them believe they could no longer be fair and impartial. Id. The court disregarded the suggestion and provided a precautionary instruction to the jury, directing the jury not to consider the transcripts. Id. at 431. This court held that “[b]y denying the reasonable request to inquire into the jurors’ states of mind, the defendants were deprived of the opportunity to meet their burden of proving actual juror bias, and were thereby denied a fair trial.” Id. at 431.

The majority distinguishes Walker because, in its view, that case involved a “highly prejudicial communication.” Maj. Op. 350. In effect, the majority determines the unauthorized communication here was not prejudicial to Taylor, notwithstanding that the district court denied *385Taylor a hearing to prove actual prejudice. Our cases involving the wrongful denial of a Remmer hearing have uniformly remanded where the extraneous statement had a potential for prejudice, because the lack of a hearing precludes a finding that the exposure was harmless. See, e.g., Davis, 177 F.3d at 557. We simply are not in a position to decide whether Juror One’s exposure to extraneous information was. harmless. Our task is to determine whether a Remmer hearing is required, which depends on whether the “alleged contact presents a likelihood of affecting the verdict.” Mack, 729 F.3d at 606. The district court assumed the statement was prejudicial, and reasonably so. (PID 2053 n.5.) This court has recognized that the term “redneck” is at least “mildly offensive.” Defoe ex rel. Defoe v. Spiva, 625 F.3d 324, 327 n. 2 (6th Cir.2010). An accusation of racism may be prejudicial, too. See Armstrong v. Shirvell, 596 F.App’x 433, 442 (6th Cir.2015) (noting that the term “racist” has a clear, well understood meaning, which is capable of being defamatory). Because the comment had a likelihood of affecting the sentence, the court should have granted Taylor’s request for a Remmer hearing.7 I would, therefore, remand for a hearing at which Taylor must be afforded a meaningful opportunity to prove bias.8

Ground II

Taylor argues that Aiken’s and Dr. Cunningham’s testimony was admissible to rebut the Government’s evidence of Taylor’s future dangerousness. The Constitution and FDPA give a defendant the right to rebut the government’s allegation of future dangerousness with relevant evidence. See Davis v. Coyle, 475 F.3d 761, 771 (6th Cir.2007) (discussing due process requirements outlined in Skipper v. South Carolina, 476 U.S. 1, 5, 106 S.Ct. 1669, 90 L.Ed.2d 1 (1986)); 18 U.S.C. § 3593(c) (“At the sentencing hearing, information may be presented as to any matter relevant to the sentence, including any mitigating or aggravating factor permitted or required to be considered.... The government and the defendant shall be permitted to rebut any information received at the hearing....”). Taylor also contends the expert testimony was independently admissible as mitigation. (Taylor Br. 110.)

The majority characterizes the Government’s allegation of future dangerousness narrowly and concludes that the excluded expert testimony “was not relevant rebuttal evidence on future dangerousness, because none of it rebutted any of the Government’s future dangerousness arguments,” because, in its view, “[a]t no point in the proceedings did the Govern*386ment seriously contend — or even intimate — that Taylor was personally dangerous.” Rather, “any such argument would have been patently absurd given Taylor’s slight stature----” Maj. Op. 862-63; see also id. at 360. The majority also quotes approvingly from the district court’s order excluding the expert testimony as irrelevant because “no rational fact finder could conclude [Taylor] is likely to personally injure someone while in custody, whether a fellow inmate or a prison staff member.” Maj. Op. 357 (quoting Mem. Op., R. 744, PID 2242.). Based on the Government’s evidence, the court concluded: “Appropriate rebuttal to the government’s evidence in this case would be information that it would be impossible for Defendant to ever communicate with anyone outside of prison for the rest of his life or that he would not be able to influence any other inmates while in prison.” Id. (quoting PID 2242).

This characterization of the evidence is unsupported by the record. The jury heard testimony regarding the escape plan, which resulted in injury to two correctional officers and in which a pipe was going to be used against prison officials. (PID 2296, 4313, 4358, 4721, 4734.) The jury learned that Taylor physically “tried to hold [an] officer down” during the escape attempt. (PID 4315, 4359, 4724.) The jury also heard that Taylor took an officer’s radio and tried to help another inmate drag an injured officer into a cell. (PID 7316.) The inmate who hatched the plan, J.R. Uhuru, stated that “Rejón did what he was supposed to do.” (Id.) Although Taylor may not have injured the officer, the jury could have inferred from the escape attempt that Taylor might physically harm another person while in prison. See United States v. Anglin, 169 Fed.Appx. 971, 975 (6th Cir.2006) (observing that “escape is a crime of violence” in the context of a prison break). Moreover, the majority’s narrow understanding of the evidence overlooks the obvious fact that led to his conviction: Taylor killed Guy Luck.9 Thus, it is flatly incorrect that “the government has introduced no evidence that [Taylor] is likely to pose a danger of personally injuring another inmate in prison or personally attacking a prison staff member.”10 (PID 2245.)

The majority’s “unrealistically limited” view of the evidence echoes the same “fallacy” the Supreme Court rejected in Kelly v. South Carolina, 534 U.S. 246, 254-55, 122 S.Ct. 726, 151 L.Ed.2d 670 (2002). In that case, the Court determined that the South Carolina Supreme Court erred “on the facts” because the evidence and arguments in the case were “flatly at odds with the view that future dangerousness was not an issue in the case.” Id. at 253, 122 S.Ct. 726 (internal quotation marks omitted). At Kelly’s trial, the prosecutor presented “evidence that Kelly took part in escape attempts and carried a shank, and that he had been caught carrying a weap*387on and planning or participating in escape attempts.” Id. (internal quotation marks omitted). Like the majority here, the state supreme court “overlooked” that this “evidence of violent behavior in prison can raise a strong implication of generalized future dangerousness.” Id. Evidence of “Kelly’s likely behavior in prison, or his proclivity to escape from it,” was evidence from which a jury could “conclude that he presents a risk .of violent behavior” and thus a future danger to others. Id. at 253-54, 122 S.Ct. 726. Moreover, the Court noted that, as the majority does here, the state court said “nothing about the evidence of crime.” Id. at 253, 122 S.Ct. 726.

With the proper understanding that “[ejvidence of future dangerousness ... is evidence with a tendency to prove dangerousness in the future,” Taylor’s experts should have been permitted to testify about the U.S. Bureau of Prisons’ (BOP) likely security arrangements to manage Taylor’s propensity for violence — whether through direct means or otherwise. Id. at 254, 122 S.Ct. 726. Dr. Cunningham’s testimony about United States Penitentiaries’ security procedures is relevant to rebut the allegation that Taylor’s earlier escape attempt, during which he attempted • to physically restrain an officer, indicated future dangerousness. Indeed, as the majority recognizes, “a district court might reasonably consider Dr. Cunningham’s summary of BOP studies as relevant evidence to rebut allegations that an individual might be directly dangerous in the future.” Maj. Op. 362. In addition, Dr. Cunningham and Aiken proffered that the BOP would monitor and control Taylor’s ability to influence other inmates and persons in the public — for example, by monitoring his communications. That testimony is relevant to rebut negative inferences from Taylor’s alleged threat against Marshall, his use of other inmates to communicate with persons outside a detention facility, and the alleged damage to MarshalPs grandmother’s home.

Due process, as well as the FDPA, allows a capital defendant to present to the jury “all possible relevant information” to rebut the future-dangerousness aggravator. Jurek v. Texas, 428 U.S. 262, 276, 96 S.Ct. 2950, 49 L.Ed.2d 929 (1976) (plurality opinion); 18 U.S.C. § 3593(c). The district court thus abused its discretion, and the error was not harmless beyond a reasonable doubt. The court instructed the jury that the Government introduced evidence pertaining to Taylor’s “involvement in an escape attempt” as evidence supporting the future-dangerousness aggravator, but drastically limited Taylor’s ability to rebut the allegation. (PID 7766.) The jxxry subsequently found the aggravating factor “tends to support imposition of the death penalty.” (PID 2288.) I would vacate Taylor’s death sentence and remand for a new sentencing hearing.11

Ground III

Under the FDPA, a party may present information relevant to establish or rebut any mitigating or aggravating factors. See 18 U.S.C. § 3593(c); Sears v. Upton, 561 U.S. 945, 950, 130 S.Ct. 3259, 177 L.Ed.2d 1025 (2010). But such information “may be excluded if its probative value is outweighed by the danger of creating unfair prejudice, confusing the issues, or misleading the jury.” 18 U.S.C. § 3593(c).

Agent James Melia testified that Inez Marshall (Marshall’s grandmother or godmother) called and told him “it appeared *388that somebody had tried to break into her residence” because the alarm had sounded, which had never happened before, and that all the windows in another of Marshall’s relative’s house “had been broken out of the house.” (PID 7299.) The allegations, as relayed by Agent Melia, came from second- and perhaps third-hand sources: Ms. Marshall and an unknown relative. Neither person testified, and the Government provided no reason for their absence. More importantly, Agent Melia did not verify the information and there is no indication whether Ms. Marshall had personal knowledge of the alleged property destruction to her relative’s house. The Government offered no evidence corroborating the alleged attempted break-in and property damage and, indeed, concedes these instances could have been mere coincidences wholly unrelated to Taylor’s prosecution.

The majority contends the statements are relevant to put Taylor’s letter to his girlfriend in context,12 to Taylor’s level of remorse, and to the future-dangerousness aggravator.13 Assuming the allegations are true, to reach the conclusion that the statements are relevant, however, one must also assume that (1) Taylor ordered the actions through an outside associate, and (2) the actions were in retaliation against Marshall (via his relatives) for his testimony. The first inference seems unlikely, given that the Government provided no evidence suggesting Taylor ordered an associate to attack Marshall’s family, despite the fact that it monitored his communications. The second inference seems similarly implausible considering Marshall’s testimony that, even after he decided to testify against Taylor, he never felt threatened by Taylor. (PID 6718.) Because these inferences are unsupported, the statements lack an apparent or demonstrated connection to Taylor. The statements, therefore, are irrelevant. Further, any probative value is outweighed by the danger of unfair prejudice.

The Government does not argue harmlessness in its brief, perhaps because it cannot prove the admission of the evidence was harmless beyond a reasonable doubt. The district court instructed the jury to consider “the harm to the home of Joey Marshall’s grandmother” as evidence supporting the future-dangerousness aggravator. By allowing the jury to consider the statements, the court in effect inferred for the jury that Taylor could somehow direct crime from prison, although there was no evidence to support such a conclusion. The jury unanimously found the future-dangerousness aggravator and determined that the factor tended to support the death penalty. Thus, the statements likely affected Taylor’s sentence.

Ground V

I agree that Taylor has not shown plain error with regard to his claim of prosecu-*389torial misconduct and, therefore, join Part V of the majority opinion. I add that another aspect of the Assistant U.S. Attorney’s rebuttal summation is troubling. This court affords prosecutors “wide latitude” during closing argument. United States v. Lawrence, 735 F.3d 385, 431 (6th Cir.2013). However, a prosecutor “may not urge jurors to identify individually with the victims.” United States v. Boyd, 640 F.3d 657, 670 (6th Cir.2011). And although a prosecuting attorney may argue that “the people in the community have the right to expect that you will do your duty,” Lawrence, 735 F.3d at 433 (quoting Hicks v. Collins, 384 F.3d 204, 219 (6th Cir.2004)), “[a] prosecutor may not urge jurors to convict a criminal defendant in order to protect community values, preserve civil order, or deter future law breaking,” United States v. Solivan, 937 F.2d 1146, 1153 (6th Cir.1991). “[A] prosecutor illicitly incites the passions and prejudices of the jury when he calls on the jury’s emotions and fears — rather than the evidence — to decide the case.” Johnson v. Bell, 525 F.3d 466, 484 (6th Cir.2008).

After analogizing Taylor to Mr. Hyde, the prosecutor argued the following to the jury:

Now, what do we do about this? What’s the plan of action? Ladies and gentlemen, society has a right to self-defense, just like the victim Mr. Luck had a right to defend himself. He fought tooth and nail to defend himself. He had every right to kill the defendant to protect himself and to defend himself, and so do you. He couldn’t finish what he had the right to start. You not only have the right but you have the obligation to finish it.
A lot of times you’ll hear people talking about the horrible things that happen in the community — “When are they going to do something about that?” Well, ladies and gentlemen, you are now they. When it comes to Rejón Taylor, you are the ones that have to do something about it. You are the ones who have the obligation to be the sheep dog that protects the sheep, and even sometimes the wolves, from the wolf.

(PID 7755-56.) Although Taylor did not object to this part of the prosecutor’s argument and does not raise an issue on appeal, the prosecutor’s argument was improper and should not be repeated.

Ground VIII

The FDPA requires a capital jury to return special findings regarding any relevant aggravating factors listed in 18 U.S.C. § 3592, and if the jury does not unanimously find a statutory aggravator, the court must impose a lawful sentence other than death. 18 U.S.C. § 3593(d). In this case, the jury found two statutory aggravators: Luck died during the commission of a kidnapping, § 3592(c)(1), and Taylor committed the murder after substantial planning and premeditation to cause Luck’s death, § 3592(c)(9). Taylor argues there is insufficient evidence to support the jury’s finding of the substantial-planning-and-premeditation aggravator. This court will uphold the jury’s verdict if, viewing the evidence in the light most favorable to the government, “any rational trier of fact could have found the existence of the aggravating factor beyond a reasonable doubt.” United States v. Lawrence, 735 F.3d 385, 415 (6th Cir.2013).

A.

Substantial planning and premeditation relate to the offense of murder, not to the offense of conviction. See United States v. Webster, 162 F.3d 308, 325 (5th Cir.1998). The Fourth Circuit defines substantial planning and premeditation as “ ‘a higher degree of planning than would have the *390words ‘planning and premeditation’ alone — i.e., more than the minimum amount sufficient to commit the offense.’ ” United States v. Jackson, 327 F.3d 273, 301 (4th Cir.2003) (citing United States v. Tipton, 90 F.3d 861, 896 (4th Cir.1996)). Similarly, the Fifth Circuit defines substantial as “a thing of high magnitude,” and substantial planning and premeditation as “requiring a considerable amount of planning preceding the killing.” United States v. Snarr, 704 F.3d 368, 392 (5th Cir.2013) (citing United States v. Flores, 63 F.3d 1342, 1373-74 (5th Cir.1995); United States v. Davis, 609 F.3d 663, 690 (5th Cir.2010)).

B.

Two years before the crime, Taylor and his codefendants began taking various persons’ mail to obtain personal and financial information. They used this information to fraudulently apply for and receive credit cards to purchase electronics, clothes, and other items. (PID 6585-86.) To gather additional personal identification information, they broke into some of the victims’ homes while the occupants were away. (PID 6586-87.) They always took with them gloves, wire cutters, and a crowbar. (PID 6589, 6592.) One victim of this scheme was Guy Luck. (PID 6591.) They did not know Luck. (PID 6591.)

While stealing Luck’s identity, Taylor learned that Luck owned a French restaurant. (PID 6594.) Marshall and Taylor visited the restaurant, although only Taylor went inside and ordered food. (PID 6595.) Marshall testified that Taylor thought the meal was expensive, which prompted them to question what Luck did with the restaurant proceeds. (PID 6598.) From following Luck, Taylor learned that every night when the restaurant closed, Luck would return home and then drive to the bank in the morning. (PID 6599.) Subsequently, Taylor and Marshall conspired to rob Luck of that money and recruited Matthews to help. (PID 6599-6600.) Both Marshall and Matthews testified that the plan was always to rob Luck and not to murder or hurt him. (PID 6649; App. 141.)

On the morning of the murder, the three went to Luck’s house in Taylor’s Impala, bringing with them the same tools they use for burglaries, including gloves and wire cutters, but also wielding loaded guns. {See PID 6601-03; App. 140.) The record is unclear regarding who supplied or obtained the guns. Marshall testified that Taylor and Matthews picked him up and the two guns were already sitting in the car. (PID 6600-01.) He added someone gave him the .38 revolver, but could not remember who, and that Matthews took the 9mm pistol. (PID 6602.) Matthews stated that Taylor and Marshall picked him up, and when he got in the car the guns were on the floor. (App. 172-73.)

According to Marshall, when they arrived at Luck’s house, Marshall and Matthews got out of the Impala and prepared to confront Luck. (PID 6602-03.) Marshall soon returned to the car because he did not think the circumstances were ideal. {Id.) After discussion, Matthews agreed to hide under Luck’s white van in the driveway and wait for Luck to exit his house. (PID 6603.) Taylor and Marshall stayed in their vehicle and rode around the neighborhood, until they noticed Matthews walking Luck back toward the house. (PID 6604.) Seeing this, Taylor jumped out of the car and ran toward the driveway; Marshall drove off, throwing his .38 revolver and gloves into the center console. {Id.)

In contrast to Marshall’s testimony, Matthews’s FBI interview states that:

Matthews pointed a gun at [Luck], grabbed the collar of his shirt, and start*391ed walking him back into the house. Marshall and Taylor drove up into the driveway, exited the Impala, and went into the house as well. All four persons went immediately into Luck’s bedroom. Matthews gave the 9mm to Taylor who stayed in the bedroom with Luck and attempted to get Luck to tell him the pin numbers for his credit cards. Luck was so nervous that he was not able to provide the numbers.
Matthews heard Taylor tell Luck, “[I]f you shut these credit cards down I’m going to come for you.” Matthews interpreted this as a threat against Luck’s life if Luck reported the cards stolen. While Taylor was in the bedroom, Marshall and Matthews searched the rest of the house. Matthews saw Taylor with Luck’s credit cards and knows Taylor made Luck turn out his front pockets where he found approximately $600 cash.
Following the search of the house and while in the bedroom Luck was still not able to provide the pin numbers for the cards Taylor stated, “We gonna take him for a ride.”

(App. 141) (Capitalizations removed.)

Eventually, Taylor and Matthews forced Luck into the back of the van: Taylor got in the driver’s seat, and Matthews got in the back to watch over Luck. (PID 6604.) 'When Marshall drove back around, the white van Was at the top of the street, and Taylor signaled for Marshall to follow him. (PID 6605.) Marshall complied with Taylor’s request, although he claims that he did not know Luck and Matthews were in the back of the van, and he had no idea what was going on. (PID 6605-06.)

With Marshall following in the Impala, Taylor got on Interstate 75 and drove away from Atlanta. (Id.) After driving 45 minutes to one hour, Taylor pulled over on to the side of the highway and got out of the van to talk with Marshall. (Id.) Taylor gave Marshall his mother’s credit card to get some gas for the Impala. (Id.) Marshall asked Taylor what he was doing and where he was going, and Taylor told him he was looking for a place to leave the van. (Id.) According to Matthews’s FBI statement, the plan was to leave Luck on the side of the road somewhere far away, so they could get away before Luck reported the robbery. (App. 142)

Before Taylor went back to the van, he allegedly told Marshall, “That’s the guy who took the warrant out on me.” (PID 6606.) Marshall knew that Taylor had a warrant in Rockdale County, and asked Taylor how he knew. (Id.) According to Marshall, Taylor responded, “I seen some papers in the house that said Rockdale County.” (Id.) When searching Luck’s house, Detective Dunn found a paper among the piles of documents on Luck’s desk that said, “The thieves identified by Matthew K. Wolfe.” (PID 4658; App. 124.) That document had the words “Rockdale County” on it. (App 124.)

After Taylor went back to the van, he and Matthews waited in the van, watching over Luck, on the side of the highway. (PID 6607.) Marshall went to the next exit, got gas, and returned to follow Taylor. (Id.) Looking for a suitable place to drop off Luck and leave the van, they drove for another 30 to 45 minutes. (Id.)

They exited off the Interstate, to a rural area in Collegedale, Tennessee. (Id.) According to Matthews, Taylor slowed the van to let Luck out and Matthews got up to unlock the van’s side door. (App. 142.) Luck took advantage of this situation and tried to escape: He fought back and jumped on Matthews’s back. (Id.) While the van was still moving, Luck supposedly grabbed at Matthews’s 9mm pistol, and they began to fight for the weapon. (Id.) Matthews pushed Luck off, and fired one *392shot at him, hitting Luck in his right arm. (Id.) However, Luck was able to get Matthews in a chokehold, and Matthews dropped the pistol, which jammed. (Id.) While trying to drive, Taylor turned around in the driver’s seat and fired three shots from the .38 revolver.14 (R. 867-Sealed, PID 4961.) The shot that proved fatal struck Luck in his upper mouth. (PID 6611-12; 4963-65.) The order in which the shots hit Luck could not be determined. (PID 4962-63.)

The van veered off the road and almost crashed into the creek. Taylor and Matthews quickly jumped out. They left their guns and a pair of gloves in the front of the van, and left Luck bleeding out, but alive, in the back. (PID 6607-09.) The .38 revolver had two live rounds left in it. (PID 6510.) A witness who had seen the van veer off the road stopped and asked if everything was fíne. (App. 142). Taylor and Matthews told the witness everything was okay, and Marshall pulled up in the Impala to pick them, up. (Id.; PID 6608.) Taylor and Matthews ran to the car; Taylor flipped down the Impala’s license plate; and they sped back toward Atlanta. (PID 6608.)

The group arrived back in Atlanta, and Marshall dropped Taylor off, then drove home with Matthews. (PID 6615.) Matthews showered and washed off the blood, and Marshall covered the Impala with a car cover. (PID 6615-16.) Taylor told Marshall he would have somebody come by and get the car later. (PID 6615.) When Matthews got out of the shower, Marshall took Matthews’s bloody clothing, which he burned the next day, and gave him new clothes. (PID 6615, 6638.) Marshall had a friend drive him and Matthews to the hospital, and later that day, Taylor, Marshall, and two friends ate at Red Lobster. (PID 6617-18, 6623.)

C.

At trial, the Government proposed two theories of Taylor’s intent. The first theory was that Taylor, without telling his code-fendants, planned all along to kill Luck. The second and more apparent theory is that Taylor originally did not intend to kill Luck when he went to Luck’s house, but after seeing the document regarding the identity theft in Rockdale County on Luck’s desk, decided then to murder Luck and thereby eliminate him as a witness.

The Government’s first theory is wholly unsupported by the record. The second theory, however, has some evidentiary support because Marshall testified that Taylor identified Luck as a possible witness against him based on a document he had seen on Luck’s desk. However, assuming Taylor intended to eliminate Luck as a witness by murdering him, he developed that intent while inside Luck’s house. Thus, any planning Taylor had completed before that point is irrelevant to whether he substantially planned and premediated the murder. What remains, then, is the nearly two-hour drive from Atlanta to Col-legedale.

Without offering a definition of the term “substantial,” or providing a single citation to authority, the majority concludes that the Government “provided sufficient evidence of planning and premeditation in this case and so Taylor’s sufficiency argument fails with respect to that aggravator.” Maj. Op. 369. But, the FDPA requires a finding of substantial planning and premeditation; mere planning and *393premeditation will not do. 18 U.S.C. § 3592(c)(9).

To support its conclusion that Taylor substantially planned and premeditated the murder, the majority cites four pieces of evidence: (1) Taylor brought guns but not masks to Luck’s home on the day of the crime, unlike in his burglaries; (2) Taylor allegedly saw the document regarding Rockdale County on Luck’s desk; (3) Taylor drove for two hours; and (4) after the crime, Taylor flipped down the license plate, was calm, and went out to eat at Red Lobster.

First, although defendants had not taken guns with them when they burglarized unoccupied homes, they had not previously planned a robbery. Moreover, the majority’s statement that “the lack of masks is evidence that this was not just another robbery attempt” is unsupported. Maj. Op. 368. The majority cites Matthews’s plea agreement, which states, “On August 6, 2003, unlike the previous burglaries, the defendants and Taylor brought guns but did not use gloves, masks, and burglary tools as they had before at Mr. Luck’s residence.” (App. 130.) But Matthews specifically crossed out the word “masks” in the plea document. (Id.) Likewise, at trial, Marshall did not testify to using masks in their burglaries. (PID 6588-90.) Second, although the plea document supports the Government’s witness-elimination theory, without more, it does not prove substantial planning and premeditation. Third, I concede the two-hour drive is significant. However, the evidence suggests Taylor killed Luck because Luck struggled with Matthews, thwarting Taylor’s plan to deposit Luck — alive—in a remote area so that they could prolong or evade detection and apprehension.15 Fourth, accepting the majority’s premise that Taylor’s actions after the shooting can evince premeditation and planning of a murder, Taylor’s post-shooting conduct does not show a considerable amount of planning of the murder.

Marshall and Taylor, together, devised a plan to rob Luck. However, both Matthews and Marshall testified there was never a plan to murder, or even hurt, Luck. After Luck began to struggle with Matthews, Matthews shot first, then Taylor shot three rounds toward the rear of the van while driving. Taylor left Luck injured but alive and with two rounds remaining in the revolver. All this suggests a lack of substantial planning. In sum, a juror could not have reasonably found beyond a reasonable doubt that Taylor substantially planned and premeditated Luck’s murder after viewing the document on Luck’s desk.

Ground XV

After oral argument in this case, the Supreme Court held the Armed Career Criminal Act’s (ACCA) residual clause unconstitutionally vague, Johnson v. United States, — U.S. -, 135 S.Ct. 2551, 192 L.Ed.2d 569 (2015), and we granted Taylor’s request to permit supplemental briefing. Taylor argues that because 18 U.S.C. § 924(c)(3)(B) — the statute providing the definition of “crime of violence” underlying his § 924(j) convictions — contains language substantially similar to the ACCA’s residual clause, Johnson renders § 924(c)(3)(B) unconstitutionally vague as well.

*394In Johnson, the Supreme Court held that two features of the ACCA’s residual clause — the use of the categorical approach and uncertainty about the level of risk required for an offense to qualify as a “violent felony” under the statute — “conspire to make it unconstitutionally vague.” 135 S.Ct. at 2557-58. In so holding, the Court found dispositive that the residual clause “ties the judicial assessment of risk to a judicially imagined ‘ordinary case’ of a crime, not to real-world facts or statutory elements,” id. at 2257, and that it “leaves uncertainty about how much risk it takes for a crime to qualify as a violent felony,” observing, “[i]t is one thing to apply an imprecise ‘serious potential risk’ standard to real-world facts; it is quite another to apply it to a judge-imagined abstraction,” id. at 2558.

Section 924(c)(3)(B)’s definition of “crime of violence” is substantially similar to the ACCA residual clause’s definition of “violent felony,”16 and suffers from the same pair of infirmities that rendered the residual clause unconstitutional. First, it requires courts to use a categorical approach, “looking to the language of the statute, rather than the particular facts” of a case to determine whether an offense constitutes a “crime of violence,” Evans v. Zych, 644 F.3d 447, 453 (6th Cir.2011), thereby tying “the judicial assessment of risk to a judicially imagined ‘ordinary case’ of a crime.” See Johnson, 135 S.Ct. at 2557. And, like the ACCA’s residual clause, § 924(c)(3)(B) requires courts to apply an imprecise “substantial risk” standard to this “judge-imagined abstraction.” See id. at 2558.

Recognizing that § 924(c)(3)(B) requires courts to use ordinary-case analysis, the majority nonetheless concludes that because § 924(c)(3)(B) is narrower than the ACCA’s residual clause, and because “much of Johnson’s analysis does not apply to § 924(c)(3)(B),” Maj. Op. 375-76, Johnson does not render § 924(c)(3)(B) unconstitutionally vague. The majority focuses on a number of distinctions between the language of § 924(c)(3)(B) and the ACCA’s residual clause that it contends remove § 924(c)(3)(B) from Johnson’s scope. See Maj. Op. 376-77. However, to ■ the extent § 924(c)(3)(B)’s language differs from the ACCA’s residual clause, the language differences do not render the provisions meaningfully different under Johnson.

A.

The majority first contends that § 924(c)(3)(B) is materially different from the ACCA’s residual clause because the language “by its nature” in § 924(c)(3)(B) — in contrast to the residual clause’s “involves conduct” — limits the court’s inquiry to the elements of the crime. Maj. Op. 376-77. However, a court’s inquiry under § 924(c)(3)(B) is no more limited to the elements of the offense *395than it was under the ACCA’s residual clause. In Johnson, the Court observed that the residual clause required courts to go “beyond deciding whether creation of risk is an element of the crime” because “unlike the part of the definition of a violent felony that asks whether the crime ‘has as an element the use ... of physical force,’ the residual clause asks whether the crime ‘involves conduct ’ that presents too much risk of physical injury.” Johnson, 135 S.Ct. at 2557. Like the ACCA, § 924(c) employs two alternative clauses— a use-of-force-as-element clause and a risk clause. And, like the ACCA residual-clause inquiry, the “by its nature” risk inquiry asks whether notwithstanding that the offense does not have the use of force as an element, the offense as usually committed presents an unacceptable risk that force will be used. The majority reasons that the “by its nature” language ties the risk inquiry to the elements of the offense, but because the risk inquiry contemplates that force is not an element — if it were, the use-of-force clause would be operative — an intrinsic feature of the risk inquiry is that looks to the hypothetical, usual case and assesses the risk of force involved in its commission. Because the risk inquiry under § 924(c)(3)(B) is untethered either to specific elements or to actual conduct, any potentially “narrowing” language in the statute narrows the risk inquiry only to the extent judges can agree on what the “ordinary” offense entails “by its nature.” And, like the ACCA’s residual clause, § 924(c)(3)(B) “offers no reliable way to choose between ... competing accounts of what an ‘ordinary’ [crime] involves” by its nature. Johnson, 135 S.Ct. at 2558. Thus, although I agree that the Johnson Court found the ACCA residual clause inquiry too broad in part because its definition of “violent felony” “did not limit a court’s inquiry to the elements of the crime,” see Maj. Op. 377, § 924(c)(3)(B) suffers from the same infirmity. The risk clause’s “by its nature” language is an alternative to the “elements” language, and thus contemplates evaluation of the “usual case.” See 18 U.S.C. § 924(c)(3); see also Johnson, 135 S.Ct. at 2557.

B.

The majority next reasons that unlike the ACCA’s residual clause, § 924(c)(3)(B) “does not allow courts to consider conduct occurring after the crime has been committed” because it requires that “physical force be used in the course of committing the offense.” Maj. Op. 376-77 (internal quotation marks omitted). There is no question that the “in the course of’ language is narrower than the residual clause, but this does not make § 924(c)(3)(B) less vague under Johnson’s reasoning.

The majority invokes Johnson’s concern that “ ‘the inclusion of burglary and extortion among the enumerated offenses preceding the residual clause confirms that’ a court could evaluate the risk of injury arising after the crime has been completed, since ‘[t]he act of making an extortionate demand or breaking and entering into someone’s home does not, in and of itself, normally cause physical injury.’ ” Maj. Op. 377 (quoting Johnson, 135 S.Ct. at 2257). Thus, the majority reasons, while the residual clause’s inquiry is open ended, the § 924(c)(3)(B) inquiry is cabined to risk during the course of the offense. But in Leocal v. Ashcroft, 543 U.S. 1, 10, 125 S.Ct. 377, 160 L.Ed.2d 271 (2004), the Supreme Court invoked burglary as a “classic example” of a crime of violence under 18 U.S.C. § 16(b), which, as the majority notes, uses the same “in the course of committing the offense” language as § 924(c)(3)(B). And this court analogized escape from a detention facility to burglary in finding the former a crime of violence under § 16(b), focusing on the risk of detection and con*396frontation, not the risk of force normally involved in the act of escaping. See United States v. Stout, 706 F.3d 704, 708-09 (6th Cir.2013).17 Thus, although § 924(c)(3)(B), like § 16(b), is narrower than the residual clause in that it focuses on “the risk that the use of physical force against another might be required in committing a crime,” Leocal, 543 U.S. at 10, 125 S.Ct. 377, the cases demonstrate that the phrase “in the course of committing the offense” has not consistently been interpreted to exclude consideration of the risk of force after the offense has technically been completed.

Further, application of the “in the course of committing the offense” limiting language does not help narrow the application of the “substantial risk” standard to an imagined ordinary case. Section 924(c)(3)(B) still suffers from the indeterminacy this idealized-offense analysis created in the residual clause. Indeed, the Supreme Court itself has taken inconsistent positions on whether the risk of force or injury from burglary arises in the course of committing the idealized offense of after. Compare Johnson, 135 S.Ct. at 2557 (“[T]he inclusion of burglary ... among the enumerated offenses preceding the residual clause confirms that the court’s task also goes beyond evaluating the chances that the physical acts that make up the crime will injure someone.”), with Leocal, 543 U.S. at 10, 125 S.Ct. 377 (“A burglary would be covered under § 16(b) ... because burglary, by its nature, involves a substantial risk that the burglar will use force against a victim in completing the crime.”). Thus, § 924(c)(3)(B)’s narrower language does not render it meaningfully different from the ACCA’s residual clause under Johnson.18

C.

The remaining distinctions the majority draws between the two statutes do not save § 924(c)(3)(B). The majority notes that unlike the ACCA’s residual clause, § 924(c)(3)(B) does not require courts to interpret the level of risk with respect to *397“four enumerated but diverse crimes.” Maj. Op. 377. However, although the Johnson Court found relevant that the enumerated offenses preceding the ACCA’s residual clause created additional confusion for courts interpreting the term “serious potential risk,” it then reiterated that the key feature that rendered the residual clause unconstitutional was “combining indeterminacy about how to measure the risk posed by a crime with indeterminacy about how much risk it takes for the crime to qualify as a violent felony.” Johnson, 135 S.Ct. at 2558. Further, although the Court stated that its holding would not affect all statutes using terms such as “substantial risk” or “unreasonable risk” in part because most such statutes do not “link[] a phrase such as ‘substantial risk’ to a confusing list of examples,” id. at 2561, it then stated:

More importantly, almost all of the cited laws require gauging the riskiness of conduct in which an individual defendant engages on a particular occasion. As a general matter, we do not doubt the constitutionality of laws that call for the application of a qualitative standard such as “substantial risk” to real-world conduct; ... The residual clause, however, requires application of the “serious potential risk” standard to an idealized ordinary case of the crime. Because “the elements necessary to determine the imaginary ideal are uncertain both in nature and degree of effect,” this abstract inquiry offers significantly less predictability than one “[t]hat deals with the actual, not with an imaginary condition other than the facts.”

Id. at 2561 (first emphasis added) (internal citations omitted). Like, the ACCA’s residual clause, § 924(c)(3)(B) assesses an idealized ordinary case, not actual facts.19 Thus, the absence of enumerated offenses preceding § 924(c)(3)(B) hardly presents an “independently compelling difference” between this statute and the ACCA under Johnson’s reasoning. The majority’s contention that the Supreme Court’s repeated attempts and failures to interpret the residual clause “militates against” applying Johnson’s reasoning to § 924(c)(3)(B) fares no better. Although this consideration bolstered the Supreme Court’s conclusion in Johnson that the residual clause was unconstitutionally vague, it was not dispos-itive.

The government argues that even if the court considers § 924(c)(3)(B) vague in some cases, we need not find it unconstitutionally vague here because some conduct — including Taylor’s — clearly qualifies as a crime of violence. However, the Supreme Court rejected a similar argument in Johnson, stating that the Court’s “holdings squarely contradict the theory that a vague provision is constitutional merely *398because there is some conduct that clearly falls within the provision’s grasp.” Johnson, 135 S.Ct. at 2561.

Finally, as the majority notes, two other circqits have addressed the issue now before the court in the context of challenges to § 16(b). Rejecting the very distinctions the majority advances, both held the statute unconstitutionally vague under Johnson. See United States v. Vivas-Ceja, 808 F.3d 719, 722-23 (7th Cir.2015); Dimaya v. Lynch, 803 F.3d 1110, 1114-20 (9th Cir.2015).

For these reasons, I would vacate Taylor’s two convictions under 18 U.S.C. § 924(j), vacate his sentence, and remand for resentencing.

. Because I conclude the district court committed reversible error regarding grounds II, III, and VIII, which requires resentencing, I do not address Taylor's arguments in Parts VI, VII, and XII.

. Taylor only objects to the failure to provide a Remmer hearing as to Juror One.

. This seems to be an acknowledgment of the likelihood of prejudice.

. At the beginning of the first day of sentencing, the court reminded the jury en masse that the ‘‘[w]eek before last I asked each of you individually some questions.” The court asked: "Will you be able to put out of your minds anything that you may have heard from the media, coworkers, neighbors, or any other source outside of the trial, in making a decision in this phase of the case? ... This would include any of the things we discussed in the individual discussions we had two weeks ago. Do you all understand that?” (PID 7235-36.)

. The majority seems more focused on the experience and good-faith of the trial judge than the dictates of Remmer. There is no question that the district judge is among the most careful and able judges, but even the best judges make mistakes, sometimes for the best of reasons.

. Although the district court questioned the jury en masse whether they could remain impartial, this court has held that a group inquiry does not satisfy Remmer's requirements; rather, at the Remmer hearing, "the defendant ] should be accorded the opportunity to question the jurors individually and under oath.” Corrado, 227 F.3d at 536-37; see also Goins v. McKeen, 605 F.2d 947, 949, 953 (6th Cir.1979) (concluding the court failed to assure the integrity of the trial because it questioned the jury en masse rather than individually and in chambers, and although the jurors who had read a prejudicial news article assured the court that they could decide the case based on the evidence). The majority distinguishes Corrado and Goins because the district court here "made a deliberate and concerted effort to investigate potential juror prejudice. Not only did the district court question the jurors in camera, one at a time, but it repeatedly asked followup questions when the need arose. The district court’s efforts here, then, were not at all like the cursory inquiries that were not adequate in Cormdo and Goins.” Maj. Op. 350. Although the district court's investigation was not cursory, it was also not complete. The court's refusal to ask Juror One whether she could remain impartial, along with its denial of Taylor's request for a Remmer hearing, *385deprived him of a meaningful opportunity to prove the juror’s impartiality.

. Moreover, the charge of racism cannot be ignored in this cross-racial capital case. As the Supreme Court stated in Turner, a case involving a capital crime of interracial violence:

Because of the range of discretion entrusted to a jury in a capital sentencing hearing, there is a unique opportunity for racial prejudice to operate but remain undetected .... More subtle, less consciously held racial attitudes could also influence a juror's decision in this case. Fear of blacks, which could easily be stirred up by the violent facts of petitioner’s crime, might incline a juror to favor the death penalty.
The risk of racial prejudice infecting a capital sentencing proceeding is especially serious in light of the complete finality of the death sentence....

476 U.S. at 35, 106 S.Ct. 1683 (plurality opinion). Our adherence to binding precedents is all the more important in a federal cross-racial capital case, as the FDPA mandates that we independently consider whether the jury authorized the sentence of death under the influence of passion or prejudice.

. I recognize that a remand for a Remmer hearing is unnecessary if we vacate Taylor’s sentence and remand for resentencing, as I conclude we should do.

. The jury specifically found that Taylor intentionally killed Luck and intentionally inflicted serious bodily injury that resulted in Luck’s death. (PID 2286.)

. The Government’s closing argument described Taylor's dangerousness in generalized terms, as well. The Government emphasized the fact of the escape attempt and Taylor's role in the plot as evidence of future dangerousness. (PID 7709-10.) In addition, the jury charge was not limited to Taylor’s ability to influence others. The special verdict form reads: "Defendant would be a danger in the future to the lives and safety of other persons, and that this factor tends to support imposition of the death penalty.” (PID 2288; see also 7766 (jury instructions).) Moreover, the court instructed the jury that the Government introduced evidence pertaining to Taylor's "involvement in an escape attempt” as evidence supporting the aggravator. (PID 7766.)

. Because the court committed reversible error when it limited Taylor's ability to rebut the Government’s sentencing argument, I do not discuss Taylor’s second argument that the excluded testimony should also have been allowed as mitigation evidence.

. The letter seems to relate to the escape attempt. While in jail, Taylor and Marshall learned that older prisoners were planning an escape. Hoping for quick freedom, both Marshall and Taylor agreed to participate. The escape did not go as planned, everyone was caught, and Uhuru blamed Marshall for botching the escape and stated that "the word is already out in the federal joint I’m at about what Joey ... did.” (PID 7315.) In a letter from jail, Taylor told his girlfriend, "Joey is in Bradley County. I never hear from him. So many people want[ ] to do something to him. Next time I go to trial, I bet he won’t testify against me. Trust me on that!” (App. 221.)

. The majority concludes the hearsay statements are reliable, although it also concludes Dr. Cunningham's testimony is unreliable because "the facts he related [came] from second, third, or fourth hand sources ... [that were] not available for questioning at trial.” Maj. Op. 359. These conclusions are inconsistent.

. Although Marshall testified he stored the .38 revolver in the Impala’s center console, Matthews stated that Taylor retrieved the weapon from Marshall during a stop in Tennessee. (App. 142.)

. I note that in the guilt phase, when given the option to find either or both that Luck's murder "was willful, deliberate, malicious, and premeditated” and/or that the murder was committed during the perpetration of a kidnapping, the jury chose only to find the latter. The jury's sentencing phase finding of substantial planning and premeditation is at odds with its refusal to so find earlier.

. Section 924(c)(3) defines "crime of violence" as:

[A]n offense that is a felony and ... (A) has as an element the use, attempted use, or threatened use of physical force against the person or property of another, or (B) that by its nature, involves a substantial risk that physical force against the person or property of another may be used in the course of committing the offense.

18 U.S.C. § 924(c)(3).

The ACCA defines "violent felony” as:

[A]ny crime punishable by imprisonment for a term exceeding one year ... that— (i) has as an element the use, attempted use, or threatened use of physical force against the person of another [ ("use of force clause”)]; or (ii) is burglary, arson, extortion, involves use of explosives [ ("enumerated offense clause”) ], or otherwise involves conduct that presents a serious potential risk of physical injury to another [(“residual clause”)].

18 U.S.C. § 924(e)(2)(B).

. Indeed, the majority in Stout did not adopt the dissent's position, which relied on many of the same distinctions the majority makes today. See 706 F.3d at 710-12 (Donald, J., dissenting).

. Additionally, to the extent Leocal recognized a distinction between language similar to the ACCA residual clause and § 16(b), it did so in the context of holding that § 16(b) does not encompass negligent or accidental conduct, such as driving under the influence of alcohol (DUI). See 543 U.S. at 10 n. 7, 125 S.Ct. 377. Later Supreme Court cases, however, made the same distinction in the ACCA context, holding that the ACCA's residual clause similarly excluded such conduct. In Begay v. United States, 553 U.S. 137, 128 S.Ct. 1581, 170 L.Ed.2d 490 (2008), the Court relied in part on Leocal to hold that DUI was not a "violent felony” under the ACCA's residual clause because it did not involve "purposeful, 'violent,' and 'aggressive' conduct.” Id. at 145-46, 128 S.Ct. 1581. Although the Supreme Court later noted in Sykes v. United States, 564 U.S. 1, 131 S.Ct. 2267, 2275-76, 180 L.Ed.2d 60 (2011), overruled by Johnson v. United States, - U.S. -, 135 S.Ct. 2551, 192 L.Ed.2d 569 (2015), that Begay's focus on "purposeful, violent, and aggressive” conduct was not required by the residual clause's text, it did so to distinguish Begay’s analysis of the strict-liability crime of DUI from the case before it, where the statute at issue had a "stringent [knowing or intentional] mens rea requirement.” Indeed, Sykes suggested that an inquiry into whether an offense requires "purposeful, violent, and aggressive” conduct would still be relevant to evaluate whether "strict liability, negligence, and recklessness crimes” constitute violent felonies under the residual clause. See id. at 2275-76; see also id. at 2289 n. 1 (Kagan, J., dissenting) (“[U]n-derstand[ing] the majority to retain the 'purposeful, violent, and aggressive’ test, but to conclude that it is 'redundant' in this case.”); United States v. Mitchell, 743 F.3d 1054, 1061 (6th Cir.2014).

. Although a fact-centric approach might make sense in the § 924(j) context because the statute addresses present rather than past conduct, which would enable a fact-finder to focus on the risk presented by the nature of a defendant’s actual conduct, case law squarely holds that the language of § 924(c) requires courts to use a categorical approach. See Leocal, 543 U.S. at 7, 125 S.Ct. 377 (observing that because [§ 16] directs our focus to the 'offense’ of convictionf,] the court must "look to the elements and the nature of the offense of conviction, rather than to the particular facts relating to [the] crime”); Evans, 644 F.3d at 453 (relying on Leocal and applying the categorical approach to § 924(c) crime-of-violence inquiry); see also United States v. McGuire, 706 F.3d 1333, 1336-37 (11th Cir.2013) (noting that the court employs a categorical approach to determine if an offense is a "crime of violence” under § 924(c)(3)(B) because of the phrase "by its nature”); United States v. Velazguez-Overa, 100 F.3d 418, 420 (5th Cir.1996) ("[T]he phrase 'by its nature’ compels a categorical approach to determining whether an offense is a crime of violence under Section 16(b).”).