USA v. Hill

Before WIENER, DENNIS, and DUNCAN, Circuit Judges. JAMES L. DENNIS, Circuit Judge: Marc Hill, Bennie Charles Phillips, Jr., Nelson Polk, and John Scott (collectively “Defendants”), in concert with Trayvees Duncan-Bush,[1] became involved in an armored car robbery at a bank automated teller machine (ATM) scheme masterminded by Redrick Batiste.[2] The scheme involved staking out ATMs to identify when armored car drivers would replenish the cash inside and then robbing the armored car at the time of delivery by shooting and killing the driver. Batiste successfully executed this murder-robbery scheme at a Wells Fargo bank ATM in Houston, Texas in 2016 with the assistance of Hill and Polk (the “Wells Fargo murder-robbery”), resulting in the death of an armored car driver. Batiste then planned a second murder-robbery at an Amegy Bank ATM in Houston (the “attempted Amegy Bank ATM murder-robbery”) with the help of all four Defendants. Acting on a tip, and after months of surveillance of Batiste, following the Wells Fargo ATM murder-robbery, law enforcement converged on the Amegy Bank ATM the day of the planned Amegy Bank ATM murder-robbery to turn the plot into a takedown. Batiste opened fire during the ambush but was shot and killed by the officers’ return fire. Law enforcement eventually arrested Hill, Polk, Scott, and Duncan-Bush at the scene and later arrested Phillips, who was not present for the planned Amegy Bank ATM murder-robbery. As the result of the Wells Fargo ATM robbery and the attempted robbery of the Amegy ATM, the surviving Defendants were charged and prosecuted for aiding and abetting robbery, attempted robbery, and aiding and abetting the use of a firearm during a crime of violence causing death of a person. In one consolidated case, after a two-week jury trial, the Defendants were each convicted on all counts. On appeal, the Defendants each raise multiple issues challenging their convictions and sentences. For the following reasons, we AFFIRM the Defendants’ convictions and sentences as to each claim. Jurisdiction This is a direct appeal from a final decision of the United States District Court for the Southern District of Texas, imposing criminal convictions and sentences, over which this court has jurisdiction under 28 U.S.C. § 1291. Defendants timely filed their notices of appeal in compliance with Fed. R. App. P. 4(b). Background Wells Fargo Murder-Robbery and Subsequent Investigation On August 29, 2016, Batiste, assisted by Hill and Polk, shot and killed an armored car driver as he was delivering approximately $120,000 to a Wells Fargo ATM. The following month, the Houston Police Department (HPD) received an anonymous tip that Batiste had been involved in the Wells Fargo murder-robbery. HPD and the FBI’s Violent Crime Task Force (the “Task Force”) investigated the tip. Special Agent Jeffrey Coughlin headed the Task Force’s investigation. Batiste’s cell phone records and cell-site locational data showed that Batiste’s phone regularly contacted the numbers associated with Hill and Polk on the day of the incident. It also revealed that all three phones were in the bank’s area on the day of the Wells-Fargo murder-robbery and in the days leading up to the murder-robbery. Attempted Amegy Bank Murder-Robbery During September and October 2016, the Task Force surveilled Batiste and observed his practice of traveling to different ATMs and banks in Houston. In October, Duncan-Bush, a jailhouse acquaintance of Phillips, joined the scheme when Phillips called Duncan-Bush to ask if he “want[ed] to make some money.” By November, it became evident to the Task Force that Batiste was targeting an Amegy Bank ATM. The Task Force had obtained court orders for the call records and cell-tower locations of Batiste and Phillips’s phones and, a wiretap of Batiste’s phone. In late November, Phillips and Duncan-Bush met with Batiste and agreed that Duncan-Bush would “grab the black bag,” containing the cash from the armored truck, and that Phillips and Duncan-Bush would split half of the cash, while Batiste would take the other half.[3] On November 30, 2016, Hill and Batiste observed the armored car’s delivery to the Amegy Bank ATM. The Task Force’s recordings revealed that Batiste called Phillips to confirm that the armored car was coming that day. Later, Batiste told Phillips that he thought about being in “savage mode” and “tak[ing] the whole truck down[.]” The same day, Batiste sent news stories about other armored car robberies to Phillips and warned him, “ [n]o talking, bragging, posting, [or] flashing[.]” Scott also joined the conspiracy that day after Batiste called and asked him if he wanted to be “ in rotation.” On December 2, Batiste called Phillips and mentioned using an AR-15 semi-automatic rifle. He told Phillips that he had had the gun’s ballistics modified in case law enforcement recovered ballistic evidence from the shooting. The next day, Batiste and Hill discussed holding a “scrimmage,” or test run, of the robbery at the Amegy Bank ATM on December 5. Phillips brought Duncan-Bush to meet Polk for the scrimmage and picked him up afterward. Batiste later called Phillips to ask whether Duncan-Bush was still willing to participate, and Phillips answered that it was “still a go.” Phone records from December 6 and 7 showed all of the Defendants in regular communication on the days before and of the planned Amegy Bank ATM murder-robbery. The investigation culminated in a government takedown of the would-be robbers on December 7, the day of the planned Amegy Bank murder-robbery. Duncan-Bush, Polk, Hill, and Scott attempted to flee from officers but were arrested. Batiste opened fire during the takedown, but the officers’ return fire hit and killed him. Phillips, who was not at the scene of the attempted robbery, was arrested that afternoon. Consolidated Prosecution of Both Cases In March 2018, a grand jury returned a four-count indictment against Hill, Polk, Scott, Phillips, and Duncan-Bush. Hill and Polk were charged with aiding and abetting Hobbs Act robbery in violation of 18 U.S.C. §§ 1951(a) and 1952 (“Count One”), and aiding and abetting the use of a firearm during a crime of violence causing the death of a person in violation of 18 U.S.C. §§ 924(c)(1)(A)(iii), (c)(3) and (j)(1) (“Count Two”), in connection with the Wells Fargo murder-robbery during which they successfully murdered and robbed an armored car driver. Hill, Polk, Scott, Phillips, and Duncan-Bush were each charged with attempted Hobbs Act robbery in violation of 18 U.S.C. § 1951(a) (“Count Three”), and aiding and abetting the discharge of a firearm during a crime of violence in violation of 18 U.S.C. § 924(c)(1)(A)(iii) and (c)(3) (“Count Four”), in connection with the Amegy Bank ATM attempted murder-robbery, which was thwarted by police and resulted in the death of coconspirator Batiste. Duncan-Bush pleaded guilty under a plea agreement to Counts Three and Four. The first attempt at trial ended abruptly after voir dire when Hill fired his counsel and requested a continuance to obtain new counsel. After a one-week trial, the jury in the second consolidated case returned guilty verdicts on all counts against each Defendant. Hill and Polk were each sentenced to two concurrent 240-month terms on Counts One and Three, followed by two consecutive life terms each on Counts Two and Four. Scott and Phillips were sentenced to 240 months on Count Three and a consecutive life term on Count Four. Defendants now appeal their convictions and sentences to this court. Discussion Shackling At trial, the court informed the parties that the U.S. Marshals Office (the “Marshals”) had evaluated the trial as having the highest level of risk and recommended that the Defendants wear leg shackles, which would be hidden from the jury’s view by a table skirt. Alternatively, the Marshals recommended using a banded electronic restraint device under each Defendant’s clothing. The Marshals based their assessment on a combination of factors, including the fact that the Defendants were charged with “premeditated, extremely violent offenses,” that the Defendants faced significant time in custody if convicted, the Defendants’ criminal histories, and the joint nature of the trial. When trial resumed after the continuance, each Defendant wore leg shackles and an electronic restraint device. Judge Werlein, who presided over the case before its transfer to Judge Hittner, had granted the Defendants’ motion not to use leg restraints notwithstanding the Marshals’ report, based on the “representation of defense counsel that they believed that the risks are not as great as the Marshal[s]” had determined. Instead, Judge Werlein had ruled that Defendants would wear electronic restraints under their clothing. However, after the continuance and transfer to Judge Hittner, Judge Hittner ordered that the Defendants wear leg shackles covered by a table skirt as the Marshals had recommended. Judge Hittner did not explain the reason for this change on the record. Hill objected to the shackling before the trial reconvened, but the court overruled these objections, given the fact that that the shackles would not be visible to the jury. However, due to disruptive behavior during the trial, on one occasion the court ordered Hill temporarily removed from the courtroom. Hill claims that during this removal, the jury saw the shackles. Thus, he contends that the district court violated his constitutional rights by shackling him in view of the jury. The Government argues that the district court did not abuse its discretion in determining that shackling was necessary given the Marshals’ assessment that Hill posed a security threat and that, even if the jury did see Hill’s shackles when he was removed from the courtroom, Hill did not present the requisite evidence to show that he was actually prejudiced as a result. This court reviews a district court’s determination to physically restrain a defendant during trial for abuse of discretion. See United States v. Maes, 961 F.3d 366, 375 (5th Cir. 2020). “A district court abuses its discretion if it bases its decision on an error of law or a clearly erroneous assessment of the evidence.” United States v. Gentry, 941 F.3d 767 (5th Cir. 2019), cert. denied sub nom. Bounds v. United States, 140 S. Ct. 2731 (2020). The Due Process Clause generally “prohibit[s] the use of physical restraints visible to the jury[.]” Deck v. Missouri, 544 U.S. 622, 629 (2005). Visible shackling can undermine the presumption of innocence, interfere with a defendant’s ability to assist in his own defense, and undermine the “dignity” of the judicial process. Id. at 630–32. But this “constitutional requirement . . . is not absolute.” Id. at 633. A trial court may exercise its discretion to determine that restraints “are justified by a state interest specific to a particular trial,” considering factors such as potential security problems and the risk of escape. Id. at 629. The Government argues that security concerns justified the court’s decision to shackle Hill. It argues that this circuit has long understood “the need to give trial courts latitude in making individualized security determinations.” United States v. Ayelotan, 917 F.3d 394, 401 (5th Cir. 2019) (internal quotation marks omitted), cert. denied, 140 S. Ct. 123 (2019). We have held that this latitude permits courts to “rely heavily on the U.S. Marshals’ advice in considering restraints.” Id. (internal quotation marks omitted). Given that the Marshals’ conclusion that the trial had the highest level of risk, the Government argues that the district court did not abuse its discretion in shackling Hill. Hill argues that “the [c]ourt did not give sufficient reasons for restraining [him] by connecting an electronic monitor under his clothes.” This court has held that even when a district court gives no reasons for shackling a defendant, those reasons may be apparent on the record when viewed in light of the specific facts of the case. See United States v. Banegas, 600 F.3d 342, 345 (5th Cir. 2010). But where the court provides no reasons and it is not apparent on the record that shackling was justified, the burden shifts to the Government to prove beyond a reasonable doubt that the error complained of did not contribute to the verdict. Id. at 346 (quoting Deck, 544 U.S. at 635). Hill thus argues that because the Government has not proven beyond a reasonable doubt that the jury did not see his restraints, he has established that his due process rights were violated. Although Hill acknowledges the Marshals’ report, he argues that this report does not make apparent on the record that the shackling was justified because the district court improperly relied on it. He asserts that “[t]he [c]ourt['s] ruling was predicated on the recommendations of law enforcement and not by an independent evidentiary assessment by the court.” But in the leading case in this area, Deck v. Missouri, 544 U.S. 622, 633 (2005), the Supreme Court held that the decision to shackle must be made based on factors specific to the trial being considered: it did not hold that the court could not rely on an assessment of the trial’s specific factors made by the U.S. Marshals. Id. (stating that a judge, in the exercise of his discretion, may shackle a defendant even in view of the jury when justified by “particular concerns . . . related to the defendant on trial” such as “special security needs or escape risks[.]“). Given well-established subsequent precedents in this circuit indicating that courts may rely heavily on the recommendation of the Marshals, Hill’s argument is not compelling. See, e.g., United States v. Ellender, 947 F.2d 748, 760 (5th Cir. 1991); United States v. Ayelotan, 917 F.3d 394, 401 (5th Cir.), cert. denied, 140 S. Ct. 123 (2019). Additionally, this court has held that “brief and inadvertent exposure to jurors of defendants in handcuffs is not so inherently prejudicial as to require a mistrial” and that in such cases “defendants bear the burden of affirmatively demonstrating prejudice, which we refused to infer from isolated incidents.” United States v. Turner, 674 F.3d 420, 435 (5th Cir. 2012) (citing United States v. Diecidue, 603 F.2d 535, 549 (5th Cir. 1979)). Here, even taking as true Hill’s assertion that the jury saw his shackles when he was removed from the courtroom, this was a brief and inadvertent exposure[4] and an isolated incident. Therefore, Hill bears the burden of demonstrating prejudice. See Turner, 674 F.3d at 435. He does not present any evidence showing that he was actually prejudiced. We thus conclude that the district court did not abuse its discretion in shackling Hill. Removal from the Courtroom During voir dire, Hill abruptly fired his attorney and requested a continuance of trial to obtain new counsel. The court reluctantly granted the motion and rescheduled trial to begin two months later. Ultimately, Hill did not retain new counsel and instead moved to proceed pro se. The court granted Hill’s motion and appointed him standby counsel. Subsequently, one morning during the trial, the court announced outside of the jury’s presence that it had been informed by the Marshals that Hill’s wife had entered the courthouse with a razor blade hidden in court clothes that she had brought for Polk.[5] The court thus barred her from entering the courthouse for the remainder of trial. This prompted an outburst from Hill, during which he repeatedly demanded that the court identify the Marshal who found the razor blade and complained of racism, general constitutional violations, and shackling. The court warned Hill that he would be removed if his behavior did not stop and allowed Hill to confer with Scott’s counsel at Scott’s counsel’s request. However, after the jury returned to the courtroom, Hill attempted to directly address the jury. The court warned Hill again that it would remove him from the courtroom if necessary, but Hill continued to protest. The court then ordered Hill removed from the courtroom and appointed his standby counsel as his lead counsel. Hill contends that the district court violated his Sixth Amendment and due process rights by temporarily removing him from the courtroom. He argues that the court acted to remove him prematurely and failed to first employ less drastic alternatives. The Government argues that Hill’s removal was justified by his disruptive conduct and that the district court is not required to use removal only as a last resort. The parties disagree as to the standard of review under which this court should review this issue. The Confrontation Clause of the Sixth Amendment states that “[i]n all criminal prosecutions, the accused shall enjoy the right . . . to be confronted with the witnesses against him . . ..”). One of the most basic of the rights guaranteed by the Confrontation Clause is the accused’s right to be present in the courtroom at every stage of his trial. Illinois v. Allen, 397 U.S. 337, 338 (1970). Federal Rule of Criminal Procedure 43 codifies this constitutional right, as well as its exception: a defendant waives the right to be present “when the court warns the defendant that it will remove [him] from the courtroom for disruptive behavior, but the defendant persists in conduct that justifies removal from the courtroom.” FED. R. CRIM. P. 43(c)(1)(C). Based on Allen and its interpretation by other courts, the Government asserts that the correct standard of review is abuse of discretion. See Allen, 397 U.S. at 343; see also United States v. Daniels, 803 F.3d 335, 350 (7th Cir. 2015); United States v. McGill, 815 F.3d 846, 900–01 (D.C. Cir. 2016) (per curiam). Hill, on the other hand, urges that the appropriate standard is “narrow discretion” based on language from United States v. Hernandez, 842 F.2d 82, 85 (5th Cir. 1988). In Hernandez, we held that the court has only “narrow discretion in deciding whether to proceed with a trial when a defendant is voluntarily in absentia . . .” Id. (quoting United States v. Benavides, 596 F.2d 137, 139 (5th Cir. 1979) (internal quotation marks omitted)). However, Hernandez articulates the standard of review for the continuing of a trial after a defendant has voluntarily left the courtroom or failed to appear altogether, not for removal of a defendant from the courtroom. Id.; see also Benavides, 596 F.2d at 139. On the other hand, where a defendant is ordered removed from the courtroom, “trial judges confronted with disruptive, contumacious, stubbornly defiant defendants must be given sufficient discretion to meet the circumstances of each case.” Allen, 397 U.S. at 343. We therefore conclude that the correct standard of review for the court’s removal of Hill is abuse of discretion. In contending that the district court removed him prematurely, thus violating his Sixth Amendment and due process rights, Hill attempts to distinguish this case from Allen. In Allen, the court removed a defendant who consistently interrupted the judge and engaged in disruptive behavior during the court proceedings. 397 U.S. 337, 339–41 (1970). After the judge had issued several warnings, Allen was removed from the courtroom. Id. at 340. The court determined he had lost his right to be present for the proceedings. Id. at 341. Hill argues that the facts which led the Supreme Court to approve of the defendant’s removal in Allen are distinguishable from this case because the defendant in Allen personally threatened the judge, and that the other cases on which the Government relies also involved “more significant, extreme, and egregious variables.” Conversely, the Government argues that the district court correctly applied Allen. Although Hill makes much of the fact that the defendant in Allen personally threatened the judge, the Court’s conclusion in Allen does not turn on that fact. Rather, the Supreme Court held that “a defendant can lose his right to be present at trial if, after he has been warned by the judge that he will be removed if he continues his disruptive behavior, he nevertheless insists on conducting himself in a manner so disorderly, disruptive, and disrespectful of the court that his trial cannot be carried on with him in the courtroom.” Allen, 397 U.S. at 343 (internal citations omitted). The Court quoted Justice Cardozo, in Snyder v. Massachusetts, 2911 U.S. 97, 106 (1934): Although mindful that courts must indulge every reasonable presumption against the loss of constitutional rights, Johnson v. Zerbst, 304 U.S. 458, 464 (1938), we explicitly hold today that a defendant can lose his right to be present at trial if, after he has been warned by the judge that he will be removed if he continues his disruptive behavior, he nevertheless insists on conducting himself in a manner so disorderly, disruptive, and disrespectful of the court that his trial cannot be carried on with him in the courtroom. Once lost, the right to be present can, of course, be reclaimed as soon as the defendant is willing to conduct himself consistently with the decorum and respect inherent in the concept of courts and judicial proceedings. Allen, at 343 (footnotes omitted). The events leading up to Hill’s removal meet this description. Moreover, Hill concedes his behavior was disruptive. Following Hill’s outburst, the court warned Hill that he would be removed if his behavior continued and allowed him to confer with Scott’s counsel at Scott’s counsel’s request. When the jury returned, Hill continued to behave disruptively and attempted to address the jury directly. At this point, the court gave Hill yet another warning before ordering his removal from the courtroom. As in Allen, the court repeatedly warned Hill that he would be removed if he did not cease behaving disruptively, yet he did not heed those warnings. Nevertheless, Hill argues that, before ordering his removal, the court should have first exhausted less extreme alternatives. But Allen does not make “removal a last resort” or require a district court to “exhaust every other possible cure” before ordering removal. United States v. Benabe, 654 F.3d 753, 770 (7th Cir. 2011); cf. Allen, 397 U.S. at 343–44 (“We think there are at least three constitutionally permissible ways for a trial judge to handle an obstreperous defendant like Allen: (1) bind and gag him, thereby keeping him present; (2) cite him for contempt; (3) take him out of the courtroom until he promises to conduct himself properly.”). In any case, the court here did attempt alternatives before removing Hill. The court explicitly warned Hill more than once to cease his disruptive conduct lest he be removed, first allowed him to confer with Scott’s counsel instead of removing him, and then removed him only after he continued to disrupt the trial in front of the jury. Further, the court allowed Hill to return to the courtroom later that same day after a recess and following standby counsel’s assertion that he would not continue his outbursts. See Allen, 397 U.S. at 343. We therefore conclude that the district court did not abuse its discretion in temporarily removing Hill from the courtroom following his outburst. Revocation of Pro Se Status When the district court temporarily removed Hill from the courtroom, it revoked his pro se status and appointed the previously designated standby counsel as lead counsel, even once Hill was permitted to return to the courtroom. Hill contends that the district court thereby improperly violated his right to self–representation. The Government argues that the district court acted within its discretion when it revoked Hill’s pro se status. We review claims concerning the right of self-representation de novo. United States v. Jones, 421 F.3d 359, 363 (5th Cir. 2005). An improper denial of the right of self-representation requires reversal without harmless error review. United States v. Majors, 328 F.3d 791, 794 (5th Cir. 2003). The right to self-representation is necessarily implied by the Sixth Amendment, but it is not absolute. See Faretta v. California, 422 U.S. 806, 818, 824 (1975). A district court “may terminate self-representation by a defendant who deliberately engages in serious and obstructionist misconduct.” Id. at 834 n.46. This court has also indicated that defendants may waive their right to self-representation via obstructionist conduct, especially if that behavior may be interpreted as a delay tactic. See, e.g., United States v. Long, 597 F.3d 720, 726–27 (5th Cir. 2010); United States v. Weast, 811 F.3d 743, 748–49 (5th Cir. 2016). The Government relies primarily on Allen to argue that the district court acted within its discretion to revoke Hill’s pro se status “following his repeated disruptive behavior and consistent refusal to comply with the court’s warnings.” It suggests that once a pro se defendant is removed from the courtroom for disruptive behavior, the appropriate procedure is for the court to revoke pro se status. See Davis v. Grant, 532 F.3d 132, 142–45 (2d Cir. 2008). In Allen, the Supreme Court found that the district court had permissibly removed the defendant from the courtroom, and that it had permissibly revoked his pro se status based on multiple incidents of disrupting the proceedings and stating that he would continue to do so, as well as threatening the judge and tearing up his attorney’s papers. 397 U.S. at 339–41. The Supreme Court rejected the notion that the Sixth Amendment right to be present at one’s own trial is absolute regardless of the defendant’s unruly or disruptive conduct. Id. at 342. Rather, “the right of self-representation is limited by the trial court’s responsibility to maintain order and safety and to prevent disruption or delay.” United States v. Vernier, 381 Fed. App’x 325, 328 (5th Cir. 2010) (unpublished) (citing Faretta, 422 U.S. 806, 834 (1975)); see also Indiana v. Edwards, 554 U.S. 164 (2008) (holding that the court did not violate the Sixth Amendment by appointing counsel against defendant’s objection where defendant was competent to stand trial but not competent to conduct trial proceedings by himself). Hill makes several arguments that the district court erred in revoking his pro se status. First, he argues that the court could have utilized standby counsel to advise Hill that his behavior was disrespectful of the court’s protocol, which it ultimately did, but not until after “the [c]ourt had already acted prematurely in terminating [his] right to self[-]representation[.]” Hill also argues that the court should have held a recess or used standby counsel to calm him down before revoking his pro se status. However, the district court did essentially attempt to mitigate the situation both ways: by allowing Scott’s counsel to confer with Hill and by taking a break in proceedings while the jury was brought back into the courtroom. Moreover, Hill’s arguments that the district court should have taken other measures before revoking his pro se status fail for the same reasons as do his arguments that the court should have taken other measures before ordering him removed from the courtroom: it tried, but Hill’s behavior did not improve. We see no abuse of discretion. Next, Hill argues that his conduct was not so extreme as in other cases in which this court has found that revocation of pro se status was permissible, citing United States v. Long, 597 F. 3d at 726–27 and Chapman v. United States, 553 F. 2d 886, 895 (5th Cir. 1977). Further, Hill argues that his conduct was not deliberatively obstructionist. See Faretta, 422 U.S. at 834 n.46; see also Chapman v. United States, 553 F.2d 886, 894 (5th Cir. 1997) (holding that a defendant’s request to represent himself at trial may be rejected if it is intended to cause delay or gain another kind of tactical advantage). Hill argues his behavior was not “an attempt to gain a strategic or tactical advantage such as delay.” Rather, he argues that it was the result of an impulsive emotional response to the removal of his wife from the courtroom. Based on Allen and our subsequent precedents, Hill’s conduct is not distinguishable from cases in which the court found revocation of pro se status permissible. See United States v. Long, 597 F.3d 720 (5th Cir. 2010); United States v. Weast, 811 F.3d 743 (5th Cir. 2016). In Long, we found that the defendant had waived his right to assert his right to self-representation at sentencing by refusing to answer the court’s questions, repeatedly asserting “Republic of Texas psychobabble” throughout the trial, and repeatedly changing his mind about firing his appointed counsel. 597 F.3d at 727 (internal quotation marks omitted). That court stated that “[g]iven Long’s previous disruptive and uncooperative conduct, the trial court may have seen [his demand to represent himself pro se] as another delay tactic.” Id. Here, similarly, given Hill’s abrupt firing of his counsel which necessitated a two-month continuance before the recommencement of trial, as well as his continual disruption of court even in the presence of the jury, the district court did not err in concluding that Hill was acting “to delay or disrupt the trial.” Weast, 811 F.3d at 749. Therefore, the district court acted within its discretion to revoke Hill’s pro se status based on his continuing disruptive conduct. Denial of a Mistrial Following Hill’s outburst and removal from the court room, Scott and Philips moved unsuccessfully to sever their trials from the other Defendants’. Instead, the court instructed the jury not to consider the outburst as evidence in the case. Scott and Hill requested that the jurors be individually polled to determine whether this instruction would cure potential prejudice due to the outburst. The court denied the request and instead questioned the jury as a group. The court asked the jury if there was any juror who could not follow its instruction to consider only the admitted evidence when rendering a verdict for each individual Defendant, and no juror raised a hand. At the conclusion of the trial, the court gave additional limiting instructions advising the jury of its duty to consider the charges and evidence against each individual Defendant separately. Phillips contends that the district court abused its discretion in denying him both a mistrial and severance. We first consider the district court’s refusal to declare a mistrial. Although Phillips did not explicitly move for a mistrial below, the Government concedes that this error is preserved on appeal because if the court had granted the motion to sever, it would have had to declare the joint proceedings a mistrial. The Government argues that the court acted within its discretion when it denied Phillips’s motion for mistrial because it gave an appropriate limiting instruction to minimize any prejudicial effect of Hill’s outburst. When the issue is preserved, as here, this court reviews the denial of a mistrial for abuse of discretion. See, e.g., United States v. Nieto, 721 F.3d 357, 369 (5th Cir. 2013). To establish an abuse of discretion, “the defendant bears the burden of showing specific and compelling prejudice that resulted in an unfair trial, and such prejudice must be of a type that against which the trial court was unable to afford protection.” United States v. Thomas, 627 F.3d 146, 157 (5th Cir. 2010) (internal quotation marks omitted). We have held that “outbursts or other disruptive actions during the course of the trial by a defendant do not, in and of themselves, justify severance” or a mistrial. United States v. Rocha, 916 F.2d 219, 229 (5th Cir. 1990). Nonetheless, “[a] district court must be mindful of the negative impact such evidence may have upon the jury and carefully consider the possible unfair prejudice against the other defendants.” Id. at 229–30. This court has long held that an appropriate limiting instruction is sufficient to prevent the threat of prejudice by evidence which is incriminating against one codefendant but not another. See, e.g., Rocha, 916 F.2d at 228–29; United States v. DeVarona, 872 F.2d 114, 121 (5th Cir. 1989); United States v. Jones, 839 F.2d 1041, 1054 (5th Cir. 1988), cert. denied, 486 U.S. 1024 (1988); United States v. Massey, 827 F.2d 995, 1004–05 (5th Cir. 1987); United States v. Hughes, 817 F.2d 268, 272–73 (5th Cir. 1987), cert. denied, 484 U.S. 858 (1987). Limiting instructions to the jury “will generally prevent actual harm to a defendant” as “jurors are presumed to follow the court’s instructions.” United States v. Richardson, 781 F.3d 237, 246 (5th Cir. 2015). The Government argues that the district court appropriately provided curative instructions to the jury in response to any potential prejudicial effects of Hill’s outburst. The Government relies on cases in which this court has held that potential prejudice resulting from one defendant’s outburst was cured by jury instructions. United States v. Stotts, 792 F.2d 1318, 1322 (5th Cir. 1986); see also Rocha, 916 F.2d at 231. The Government argues that, as in those cases, the district court here acted within its discretion to deny a mistrial because, following Hill’s outburst, it gave detailed instructions on multiple occasions for the jury to disregard the disruption. Phillips argues that the prejudicial effect of Hill’s outburst required a mistrial be declared because it “created a unique and extreme circumstance” that could not be cured by limiting instructions. Phillips relies on Braswell v. United States, in which this court did hold that prejudice to defendants due to codefendants’ outbursts justified a mistrial. 200 F.2d 597, 602 (5th Cir. 1952). Further, whereas this court has stated that general assertions not pointing to “specific events that caused substantial prejudice” are insufficient, United States v. Smith, 895 F.3d 410, 416 (5th Cir. 2018), cert. denied sub nom. Washington v. United States, 139 S. Ct. 495 (2018), Phillips argues that, here, he has pointed to very specific instances of prejudicial outbursts. Although we recognized in Braswell that a disruption by a codefendant may result in incurable prejudice, on review of the facts, the disruption in Braswell was much more extreme than in this case. Braswell v. United States, 200 F.2d 597, 602 (5th Cir. 1952). In Braswell, two codefendants had assaulted a U.S. Marshal during the trial and another defendant had to be forcibly restrained to prevent her from taking pills, during which she bit a police officer. Id. In comparison, we have held under similar and more extreme circumstances than those presented here that jury instructions to disregard the incident cured any possible prejudice from the codefendant’s outburst. See United States v. Stotts, 792 F.2d 1318, 1322 (5th Cir. 1986) (finding prejudice to be effectively cured by jury instructions to disregard a codefendant’s outburst during which he was removed from the courtroom after an “altercation” with a Marshal); Rocha, 916 F.2d at 231 (finding prejudice to be effectively cured by jury instructions to disregard a codefendant’s outburst during which he made a death threat to a witness during that witness’s testimony). We therefore hold that Hill’s outburst falls short of the rare circumstances in which a codefendant’s disruption results in incurable prejudice such that a mistrial is required. Denial of Motions to Sever Next, we consider Phillips’s severance motion. Phillips contends that the district court abused its discretion in denying his motion for severance, which he filed after Hill’s outburst and removal from the courtroom. The Government argues that the district court acted within its discretion when it denied Phillips’s request to sever. This court reviews a district court’s “decision not to sever under the exceedingly deferential abuse of discretion standard.” United States v. Daniel, 933 F.3d 370, 380 (5th Cir. 2019) (internal quotation marks omitted). Moreover, we will not reverse a district court’s decision not to sever unless the defendant establishes “clear, specific and compelling prejudice that resulted in an unfair trial.” United States v. Huntsberry, 956 F.3d 270, 287 (5th Cir. 2020) (internal quotation marks omitted). On top of the abuse of discretion standard, a defendant challenging the court’s denial of his request to sever also faces a second burden of precedent, which “does not reflect a liberal attitude toward severance.” Daniel, 933 F.3d at 380. Rather, “[t]o promote judicial economy and the interests of justice,” there is a strong preference in the federal system for joint trials of defendants indicted together. Id.; see also Zafiro v. United States, 506 U.S. 534, 540 (1993). To overcome this high burden, the defendant must show a “specific and compelling prejudice” resulting from the joint trial. United States v. Owens, 683 F.3d 93, 98 (5th Cir. 2012). The defendant must also show that he was not adequately protected from this prejudice by limiting instructions to the jury, id., and that this prejudice “outweighed the government’s interest in economy of judicial administration[.]” Daniel, 933 F.3d at 380. Additionally, it is not enough for a defendant to “alleg[e] a spillover effect[,] whereby the jury imputes the defendant’s guilt based on evidence presented against his codefendants[.]” United States v. Reed, 908 F.3d 102, 114 (5th Cir. 2018) (internal quotation marks omitted), cert. denied, 139 S. Ct. 2655 (2019). Rather, “severance is required on the basis of a disparity in the evidence only in the most extreme cases.” Owens, 683 F.3d at 100 (emphasis in original) (internal quotation marks omitted). Even in cases involving a high risk of prejudice, limiting instructions will often suffice to cure this risk. Id. at 381. The Federal Rules of Criminal Procedure do not require severance based on prejudice, but provide that the court may sever or “provide any other relief that justice requires.” FED. R. CRIM. P. 14(a); see also Zafiro, 506 U.S. at 539 (explaining that Rule 14 “leaves the tailoring of the relief to be granted, if any, to the district court’s sound judgment.”). To overcome the presumption that juries “follow the instructions given to them by the district court,” a defendant “must identify specific instances of prejudice unremedied by limiting instructions.” Daniel, 933 F.3d at 381. Further, a “conclusory assertion” that the jury was unable to follow limiting instructions is insufficient. Reed, 908 F.3d at 114. Phillips argues that the prejudice against him was specific and compelling enough that severance was required. He contends that while he was charged under the same superseding indictments as his codefendants, he was not involved in or charged with the death of the armored car driver during the Wells Fargo murder-robbery, and thus that the evidence presented to support that count was severely prejudicial to him. The Government argues that the district court acted within its discretion in declining to order severance because, first, the Wells Fargo murder-robbery and the Amegy Bank attempted murder-robbery were “so completely intertwined,” and, second, the district court gave strong limiting instructions throughout the trial to minimize the risk of prejudice. Phillips urges that the facts here resemble cases in which we have held that the district court abused its discretion in denying appellant’s request for severance. See, e.g., United States v. McRae, 702 F.3d 806, 828 (5th Cir. 2012). However, these cases are distinguishable. In McRae, this court held that the district court abused its discretion in declining to sever the case of one defendant, Warren, from those of his codefendants. 702 F.3d 806 (5th Cir. 2012). There, Warren was charged only with depriving a victim, Glover, of his right to be free from the use of unreasonable force by a law enforcement officer, and carrying, using, and discharging a firearm in furtherance of a felony crime of violence resulting in death. His codefendants, on the other hand, were charged additionally under civil rights statutes for beating two men, burning one of their cars, and burning Glover’s body. Id. at 824. Warren’s codefendants were also charged with obstruction of justice for interference with the investigation into these crimes, and with the use of fire to commit civil rights deprivations and obstructions, with preparing and submitting a false narrative with intent to obstruct the investigation of the Glover shooting, and with making false statements to a federal grand jury. Id. This court found there that the district court had erred by refusing to sever Warren’s trial. Id. at 842. In making this determination, we emphasized that if Warren had been tried alone the trial would have lasted approximately three days, whereas there he endured a month-long trial saddled by prejudicial evidence and testimony unrelated to his charges. Id. at 825–26. We have indicated that the McRae decision was narrow and based on the facts presented in that case. For example, in United States v. Reed, we stated, Steven Reed points to our decision in [McRae] where we reversed a district court’s refusal to sever one police officer’s officer-involved shooting trial from the trial of a set of other police officers who separately attempted to cover up the shooting. Unlike in McRae, the evidence presented against Walter Reed on the counts only pertaining to him (the tax return, mail fraud, and certain wire fraud counts) was not so inflammatory that the jury would find it highly difficult to dissociate it from Steven Reed’s conduct. Further, the charge and evidence against Steven Reed was significantly related to the charge and evidence against Walter Reed on the campaign funds counts, whereas in McRae, two sets of defendants were effectively being tried for two completely different offenses and the only link was that one offense was the “catalyst” for the other. 903 F.3d 102, 114 n.40 (5th Cir. 2018); see also United States v. Ledezma-Cepeda, 894 F.3d 686 (5th Cir. 2018) (distinguishing McRae on the grounds that, in McRae, Warren was not a member of the conspiracy and had committed crimes qualitatively less severe than those of his codefendants). In comparison, here, the evidence presented against Phillips’s codefendants alone “was not so inflammatory that the jury would find it highly difficult to dissociate it from” Phillips’s conduct. Reed, 903 F.3d at 114 n.40. As in Reed, “the charge and evidence against [Phillips] was significantly related to the charge and evidence” against his codefendants “whereas in McRae, two sets of defendants were effectively being tried for two completely different offenses and the only link was that one offense was the “catalyst” for the other.” Id. Although Phillips was not charged with Counts 1 and 2 regarding the Wells Fargo robbery, and makes much of the fact that that robbery resulted in the death of an armored truck driver, Phillips was charged for the Amegy Bank robbery, during which Batiste’s death occurred and which involved the planned murder of another armored truck driver. Under those circumstances, the evidence presented against the other Defendants on Counts 1 and 2 was not so much more inflammatory than the conduct for which Phillips was charged that “jury would find it highly difficult to dissociate it from” Phillips’s own conduct. Reed, 903 F.3d at 114 n.40. The charges against Phillips do not differ dramatically from those against his codefendants. See id.; see also United States v. Erwin, 793 F.2d 656, 666 (5th Cir. 1986) (finding error in refusing to sever as to one defendant against whom the charges were “only peripherally” related to those against the other defendants). Therefore, we hold that the district court did not abuse its discretion in denying Phillips’s motion for severance. Agent Coughlin’s Testimony Regarding Coded Language At trial, the Government’s case-in-chief began with Agent Coughlin, whose testimony focused in part on cell phone evidence from the wiretapping of Batiste’s phone. Agent Coughlin testified that the investigation had occupied “75 to 80 percent of [his] time[,]” and that he had spent “ a massive amount of time” reviewing all the evidence. When the Government played Batiste’s wiretapped calls, Agent Coughlin frequently provided interpretations of any coded language. For example, he explained that Batiste’s reference to “savage mode” meant executing the robbery while armored car guards moved the money from a broken armored truck to a second truck. Phillips argues that the district court erred by allowing Agent Coughlin to provide lay-opinion testimony regarding his interpretation of coded language in the wiretapped phone calls. The Government argues that the district court did not err, much less commit reversible plain error, by allowing Agent Coughlin’s lay testimony about the wiretapped phone calls. The parties debate the applicable standard of review. Phillips argues that the issue should be reviewed for abuse of discretion. The Government contends that the issue should be reviewed for plain error. This court reviews “preserved objections regarding the admission of expert or lay testimony for abuse of discretion, subject to harmless error analysis.” United States v. Haines, 803 F.3d 713, 726 (5th Cir. 2015). “Unpreserved errors of the same variety are reviewed for plain error.” Maes, 961 F.3d at 372. “To be considered preserved for appeal, a defendant’s objection to a district court’s ruling must be on the specific grounds raised below.” Id. Phillips argues that the standard of review for the admissibility of the lay-opinion testimony is abuse of discretion because that is the applicable standard of review on appeal for the admissibility of evidence. United States v. Westmoreland, 841 F.2d 572, 578 (5th Cir. 1988), cert. denied, 488 U.S. 820 (1988); United States v. Stephenson, 887 F.2d 57, 59 (5th Cir. 1989), cert. denied, 493 U.S. 1086 (1990). However, as the Government points out, this case presents two wrinkles. First, Phillips did not object to Coughlin’s testimony: his codefendant Scott did. A defendant typically “must bring his own objections to preserve them.” United States v. Evans, 892 F.3d 692, 711 n.1 (5th Cir. 2018). However, we have sometimes considered an evidentiary objection by a codefendant “sufficient to invoke the abuse of discretion standard[.]” United States v. Sanchez-Sotelo, 8 F.3d 202, 210 (5th Cir. 1993); see also United States v. Westbrook, 119 F.3d 1176, 1185 (5th Cir. 1997); but see United States v. Belanger, 890 F.3d 13, 27 (1st Cir. 2018) (reviewing argument concerning wiretap evidence for plain error when only a codefendant objected). Second, even assuming Scott’s objection was adequate to preserve the issue for abuse of discretion review, the Government argues that it should only extend to the “specific grounds raised below” by Scott. Maes, 961 F.3d at 372 (internal quotation marks omitted). The Government contends that Scott’s objections were not that Agent Coughlin was unqualified to provide lay testimony on the meaning of the coded language in the wiretapped calls, but instead that he challenged only Coughlin’s testimony regarding two specific calls. Thus, the Government contends, only a challenge to Coughlin’s testimony regarding those two calls would be preserved for abuse of discretion review, and the rest would be reviewed for plain error. Because we find that Phillips’s claim fails under either abuse of discretion or plain error review, we go forward applying the less stringent abuse of discretion review. Federal Rule of Evidence 701 provides that a witness may offer lay opinion testimony when “it has the effect of describing something that the jurors could not otherwise experience for themselves by drawing upon the witness’s sensory and experiential observations that were made as a first-hand witness to a particular event.” United States v. Haines, 803 F.3d 713, 733 (5th Cir. 2015) (cleaned up). By contrast, a witness’s “[t]estimony on topics that the jury is fully capable of determining for itself is not ‘helpful to clearly understanding the witness’s testimony,’ and therefore is inadmissible under Rule 701.” Id. (citing, quoting FED. R. EVID. 701). As the Government points out, this court has consistently held that law enforcement agents may “draw upon their familiarity with a particular case . . . to provide lay opinion testimony regarding the meaning of specific words and terms used by the particular defendants in the case.” United States v. Staggers, 961 F.3d 745, 761 (5th Cir. 2020) (internal quotation marks omitted), cert. denied, 141 S. Ct. 388; accord, e.g., Haines, 803 F.3d at 729; United States v. Akins, 746 F.3d 590, 599–600 (5th Cir. 2014); United States v. El-Mezain, 664 F.3d 467, 514 (5th Cir. 2011). “[E]xplaining the meanings of terms as used in the conversations and documents, as well as the relationships between the people the agent is investigating, provides the jury with relevant factual information about the investigation.” Haines, 803 F.3d at 729 (cleaned up). The Government argues that, as in those cases, the district court here properly allowed Agent Coughlin’s lay testimony of his interpretation of the calls because his participation in the case was extensive. See, e.g., Staggers, 961 F.3d at 761 (summarizing a case agent’s extensive involvement in the investigation); Akins, 746 F.3d at 599–600 (same). Coughlin not only led the investigation from the start, but he also spent “75 to 80 percent of [his] time” at work on the case. Coughlin “had much more insight into the meaning of the code words than did the jury.” United States v. Macedo-Flores, 788 F.3d 181, 192 (5th Cir. 2015) (approving coded language testimony). Coughlin was therefore qualified to provide his opinion “regarding the meaning of specific words and terms used by the particular defendants in the case.” Staggers, 961 F.3d at 761 (internal quotation marks omitted). Phillips argues that Agent Coughlin’s testimony usurped the function of the jury to draw inferences on its own from the evidence presented. He cites only one precedential case[6] to support the argument that the admission of Coughlin’s testimony was improper: United States v. Haines, 803 F.3d 713 (5th Cir 2015). Haines fails to help Phillips. There, a DEA agent testified to his interpretations of jargon in intercepted calls to prove a drug conspiracy. Id. at 713. We concluded that the agent’s testimony was admitted in error because “it went beyond [the agent]‘s expertise and personal knowledge of the investigation and instead ventured into speculation, usurping the jury’s function, which is to draw its own inferences from the evidence presented.” Id. at 734. But in Haines we made a distinction between the kind of lay testimony as to the meaning of coded words based on an agent’s “experiential observations[,]” see Haines, 803 F.3d at 733, which we found permissible, and testifying as to the meaning of common words such as, in that case, “what,” “she,” “that,” and “stuff,” which we found impermissible. See also United States v. Peoples, 250 F.3d 630, 639–40 (8th Cir. 2001) (making the same distinction). As the Government points out, Phillips’s argument does not account for the different holdings for these two categories. Here, Coughlin’s testimony falls into the first, permissible category. The coded meanings about which Coughlin testified were not as to common words, but rather to opaque terms and phrases such as “the commissary is coming,” “savage mode,” “hellos,” and “African devil.” This court has approved coded-language testimony under similar circumstances. See Haines, 803 F.3d at 729 (proper for agent to opine that “the phrase ‘I’ll be up there’ is a reference to Houston, Texas”); Staggers, 961 F.3d at 761 (proper for agent to opine “that the terms ‘gator meat’ and ‘alligator’” referred to heroin). Coughlin’s testimony therefore did not, as Phillips argues, impermissibly usurp the function of the jury. Confrontation Clause Agent Coughlin’s testimony also concerned reports of information extracted from Defendants’ cellphones. However, rather than the full, mechanically extracted reports, Coughlin testified to versions of the extraction reports that he had himself edited down to those portions he deemed relevant. Polk, Scott, and Hill raised Sixth Amendment objections to this, asserting that, because Coughlin did not personally extract the reports from their cell phones or observe the extraction, his testimony violated the Sixth Amendment’s Confrontation Clause. The court overruled the objections, accepting the Government’s argument that the reports were not “opinion piece[s]” in which someone was “ evaluating the evidence[.]“ Polk and Scott argue on appeal that the district court violated their Sixth Amendment rights under the Confrontation Clause by allowing Agent Coughlin to testify concerning data reports which were extracted from Defendants’ cell phones. The Government argues that the cell-phone extraction reports were not testimonial statements triggering the Confrontation Clause because the reports are raw, machine produced data that contained no independent analysis or opinion. “This court reviews de novo a timely Confrontation Clause objection, subject to harmless error analysis.” United States v. Morgan, 505 F.3d 332, 338 (5th Cir. 2007) (per curiam). But when the defendant’s objection is untimely, this court’s review is for plain error. United States v. Martinez-Rios, 595 F.3d 581, 584 (5th Cir. 2010) (per curiam). The parties disagree about whether the Defendants’ objections were timely. The Government argues, based on the Southern District of Texas Criminal Local Rules, that Defendants were required to make any objection to exhibits at least seven days before trial, and that failure to object in writing pretrial “concedes authenticity.” S.D. Tex. Crim. L.R. 55.2. Thus, it argues that the Defendants’ objections made during Coughlin’s testimony were untimely. However, as Scott points out, Judge Werlein had specifically stated that he would rule on any objections to exhibits at the time they were offered. And Judge Hittner, once the case was transferred to him, stated that all of Judge Werlein’s former rulings remained in effect. Arguably, then, this relieved Defendants of the requirement to bring objections to exhibits in writing at least seven days before trial. We thus proceed on the assumption that Judge Werline’s rulings remained in effect and that de novo review is the correct standard of review. The Confrontation Clause of the Sixth Amendment, in pertinent part, provides that “[i]n all criminal prosecutions, the accused shall enjoy the right . . . to be confronted with the witnesses against him.” U.S. CONST. AMEND. VI. In Crawford v. Washington, 541 U.S. 36, 68 (2004), the Supreme Court held that fidelity to the Confrontation Clause permitted admission of “[t]estimonial statements of witnesses absent from trial . . . only where the declarant is unavailable, and only where the defendant has had a prior opportunity to cross-examine.” Id. at 59; see also Michigan v. Bryant, 562 U.S. 344, 354 (2011) (“[F]or testimonial evidence to be admissible, the Sixth Amendment ‘demands what the common law required: unavailability [of the witness] and a prior opportunity for cross-examination.’” (quoting Crawford, 541 U.S. at 68)). In Melendez–Diaz, relying on Crawford‘s rationale, the Court refused to create a “forensic evidence” exception to this rule. 557 U.S. 305, 317–21. There, the Court held that an analyst’s certification prepared in connection with a criminal investigation or prosecution was “testimonial,” and therefore within the compass of the Confrontation Clause. Id. at 321–324. Applying Melendez-Diaz, the Supreme Court held that a forensic analyst who had not performed or observed a blood-alcohol test could not testify to the forensic report certifying the test’s result under the Confrontation Clause. Bullcoming v. New Mexico, 564 U.S. 652, 662 (2011). But on the other hand, on plain error review, this court has found no error in district courts admitting reports containing only “raw, machine-produced data[;]” in those cases, GPS cellphone tracking reports. See United States v. Waguespack, 935 F.3d 322, 333–34 (5th Cir. 2019), cert. denied, 140 S. Ct. 827 (2020); United States v. Ballestros, 751 F.App’x 579, 579–80 (5th Cir. 2019) (unpublished), cert. denied, 139 S. Ct. 2706 (2019). In so doing, we have explained that multiple other circuits have also held that “machine statements aren’t hearsay.” United States v. Lizarraga-Tirado, 789 F.3d 1107, 1110 (9th Cir. 2015) (satellite images with machine generated location markers); United States v. Lamons, 532 F.3d 1251, 1263 (11th Cir.2008) (cell phone call and billing records); United States v. Moon, 512 F.3d 359, 362 (7th Cir. 2008) (raw drug test data; “The report has two kinds of information: the readings taken from the instruments, and [the witness's] conclusion that these readings mean that the tested substance was cocaine.”); United States v. Washington, 498 F.3d 225, 230 (4th Cir. 2007) (raw drug test data); United States v. Hamilton, 413 F.3d 1138, 1142 (10th Cir. 2005) (computer generated ‘header’ information); United States v. Khorozian, 333 F.3d 498, 506 (3d Cir. 2003) (same). The Government argues that, applying those principles here, the cell-phone extraction reports that Agent Coughlin testified about were not testimonial statements triggering the Confrontation Clause. Rather, the Government asserts that, unlike the forensic reports at issue in Bullcoming and Melendez-Diaz, these reports are raw, machine produced data that contained no independent analysis or opinion. In the alternative, if the reports are testimonial, the Government argues that Coughlin was in fact the correct witness to testify to them as he was the one who curated the tens of thousands of pages of data extracted from the cellphones into the excerpted versions containing only the information which Coughlin deemed relevant from which he testified. Polk and Scott aver that the cellphone extraction reports are testimonial; thus, that Coughlin’s testimony about the extraction reports violated their Confrontation Clause rights. They point out that, as in Bullcoming, Coughlin did not participate in or observe the creation of the extraction reports. Further, the Defendants argue that the extraction reports were similar to the forensic laboratory report in Bullcoming and were thus testimonial evidence subject to the Confrontation Clause. We agree with the Government that the extraction reports at issue here were non-testimonial, raw machine created data. Key differences exist between test reports generated by a person’s analysis and test reports which are the result of machine analysis. This distinction has been illustrated by Bullcoming and its impact on the progeny of the Seventh Circuits’ Moon, 512 F.3d at 362, and the Fourth Circuits’ Washington. 498 F.3d at 230. As the Fourth Circuit pointed out in United States v. Summers, the Supreme Court in Bullcoming emphasized that the report in question there “contained not only raw, machine-produced data, but also representations relating to past events and human actions[,]” e.g., the validity of the analysis or the integrity of the sample. 666 F.3d 192, 199 (4th Cir. 2011) (emphasis original) (cleaned up) (citing Bullcoming, 564 U.S. at 660). Albeit on plain error review, this court has made similar holdings, see Waguespack, 935 F.3d 322 (5th Cir. 2019), following the logic of Supreme Court precedent in Melendez-Diaz, 557 U.S. at 311, and Bullcoming, 564 U.S. at 662, in which the Court emphasized that the reports in question were analyzed by a person and were not “only machine-generated results, such as a printout from a gas chromatograph.” Bullcoming, 564 U.S. at 673 (Sotomayor, concurring in part). Here, the raw cellphone extraction reports contained “only machine-generated results,” and were thus non-testimonial. Even if we were to construe the curated extraction reports which were actually admitted into evidence and testified about by Coughlin as testimonial, Coughlin would be the correct person to testify about those reports because he created them from the raw data. Scott argues that this holding is akin to allowing the Government to introduce an “excerpt of an autopsy report through a witness, claiming that the witness is the declarant of those excerpts from the autopsy report since he created the excerpt[.]” But this argument assumes that the underlying report being excerpted is itself testimonial. We therefore hold that the district court did not err in allowing Coughlin to testify to the extraction reports he had excerpted from the full, raw machine-generated reports of Defendants’ cellphone data. Ex Parte Contact with a Juror At one point during the trial, the Government notified the court that it had learned of an incident in which someone from the courtroom gallery followed a juror out of the courthouse and called that juror by name. The court confirmed with the Marshals that the unidentified person was not someone on the witness list.[7] Polk asked the court to identify the juror and the court refused, but the court did agree to conduct a general inquiry and requested that defense counsel collaborate on a limiting instruction to the jury. After discussion, Polk stated that the Defendants were “concerned about questioning the jury and poisoning the jurors with information that they don’t already have, or they may not even be aware of.” However, the court again declined to identify the juror, and after further discussion, Scott told the court that the Defendants wanted to give a jury instruction before excusing the jury at the end of the day. Hill provided the proposed jury instruction,[8] to which each Defendant and the Government agreed. At the end of the day, the court gave the instruction, and, after giving the jurors the opportunity to ask questions or express any problems with the instruction, the case manager stated that no juror had expressed concern about the instruction. The court stated that, in that case, it did not need to call any jurors back to discuss it, and none of the parties objected. Hill, Polk, and Scott claim that the district court abused its discretion by failing to adequately respond to this incident of alleged ex parte contact with a juror. The Government responds that the Defendants waived this argument via their conduct at trial. While we disagree that the argument has been waived, we hold that the Defendants’ argument fails on its merits. This court reviews a district court’s decision “in handling complaints of outside influence on the jury” for abuse of discretion. United States v. Sotelo, 97 F.3d 782, 794 (5th Cir. 1996). “The district court must balance the probable harm resulting from the emphasis that a particular mode of inquiry would place upon the misconduct and the disruption occasioned by such an inquiry against the likely extent and gravity of the prejudice generated by the misconduct.” Id. We “accord broad discretion to the trial court in these matters[,]” recognizing the district court’s unique ability to evaluate the “ mood and predilections of the jury[.]” Id. The Government argues that the Defendants waived this argument by formulating and agreeing to the jury instruction given in response to the ex parte contact. “A waiver occurs by an affirmative choice by the defendant to forego any remedy available to him, presumably for real or perceived benefits resulting from the waiver.” United States v. Richard, 901 F.3d 514, 517 (5th Cir. 2018) (internal quotation marks omitted). However, the cases the Government cites to support this argument found waiver where a defendant affirmatively agreed to a jury instruction and then sought to claim error based on the instruction itself. See United States v. LeBeau, 949 F.3d 334, 342 (7th Cir. 2020); cert. denied, 141 S. Ct. 261; United States v. Feldman, 931 F.3d 1245, 1260 (11th Cir. 2019), cert. denied, 140 S. Ct. 2658 (2020). The Government also relies on an unpublished case from this circuit which found waiver where a defendant sought to challenge the court’s resolution of an issue when he had explicitly agreed to the decided course of action in a prior proceeding. United States v. Hoover, 664 F. App’x 363, 366 (5th Cir. 2016) (unpublished). We disagree that the Defendants have waived this issue. Unlike in the cases on which the Government relies, the Defendants here did not affirmatively agree with the district court’s course of action in attempting to rectify the ex parte contact with a jury instruction; instead, once the court determined that a jury instruction would suffice to rectify the alleged ex parte contact, the Defendants agreed to the wording of the instruction itself, which they do not challenge here. The argument that the Defendants seek to raise— that the district court did not sufficiently inquire into the alleged ex parte contact before determining that a jury instruction would be sufficient to cure any resultant prejudice— was therefore not waived. Nonetheless, we agree with the Government that the Defendants’ argument fails on the merits. We afford broad discretion to district courts to tailor the appropriate response to incidents like this, trusting them, as the courts of first impression, to “balance the probable harm resulting from the emphasis that a particular mode of inquiry would place upon the misconduct and the disruption occasioned by such an inquiry against the likely extent and gravity of the prejudice generated by the misconduct.” Sotelo, 97 F.3d at 794. The Government contends that the district court’s response to the alleged ex parte contact was wholly within its discretion. By adopting a neutral cautionary instruction, the Government urges that the district court acted well within the court’s “broad discretion to fashion an investigation.” Sotelo, 97 F.3d at 797. Moreover, as the Government points out, the Defendants themselves recognized the risk that a formal investigation of this incident might itself cause prejudice by providing the jurors with information they did not already have. Defendants argue that the district court abused its discretion in failing to conduct a sufficient inquiry into the ex parte contact. Polk contends that “the nature, circumstances, prejudicial impact on the case, and how it affected the jury was not investigated much less determined.” Hill argues that the court should have called potential witnesses to determine the prejudicial impact of the contact. However, the Defendants’ arguments ultimately amount to a disagreement with the mode of inquiry chosen by the court to investigate and address the ex parte contact. A requirement like the one Defendants propose, that a district court inquire in a specific way into an allegation of ex parte contact, does not comport with the broad discretion afforded to district courts to individually tailor effective mitigation of such incidents. We therefore hold that the district court did not abuse its discretion by insufficiently inquiring into the allegation of ex parte contact. Sufficiency of the Evidence Phillips next contends that the evidence was insufficient to support either of his convictions of Count Three, attempted Hobbs Act Robbery, 18 U.S.C. § 1951(a), or Count Four, aiding and abetting the discharge of a firearm during a crime of violence. 18 U.S.C. § 924(c)(1)(A)(iii); (c)(3). The Government argues that, viewing the evidence in the light most favorable to the verdict, there was sufficient evidence to support Phillips’s convictions. When a defendant preserves a challenge to the sufficiency of the evidence, this court’s review is de novo. See, e.g., United States v. Dailey, 868 F.3d 322, 327 (5th Cir. 2017). Sufficient evidence supports a jury’s verdict so long as “a rational trier of fact could have found the elements of the crime beyond a reasonable doubt.” Dailey, 868 F.3d at 327. Sufficiency review is “highly deferential” to the jury’s determination of guilt. United States v. Zamora-Salazar, 860 F.3d 826, 831 (5th Cir. 2017). This court may not reweigh the evidence, nor second-guess “[c]redibility choices that support the jury’s verdict[,]” id. at 832; rather, it must view all evidence, reasonable inferences, and credibility choices in the light most favorable to that verdict. See, e.g., Dailey, 868 F.3d at 327. Count Three There are two elements of a Hobbs Act violation: “(1) robbery, extortion, or an attempt or conspiracy to rob or extort (2) that affects commerce.” United States v. Avalos-Sanchez, 975 F.3d 436, 440 (5th Cir. 2020) (footnotes omitted). To be convicted of attempt, “the evidence must show the defendant (1) acted with the culpability required to commit the underlying substantive offense, and (2) took a substantial step toward its commission.” United States v. McGee, 821 F.3d 644, 647 (5th Cir. 2016) (internal quotation marks omitted). A defendant’s “mere preparation” does not meet the substantial-step requirement. See, e.g., United States v. Howard, 766 F.3d 414, 419 (5th Cir. 2014). But a substantial step “is less than the last act necessary before the crime is in fact committed[;]” it simply requires “ conduct that strongly corroborates the firmness of the defendant’s criminal intent.” Id. (internal quotation marks omitted). This requirement “prevents the conviction of persons engaged in innocent acts on the basis of a mens rea proved through speculative inferences, unreliable forms of testimony, and past criminal conduct.” United States v. Oviedo, 525 F.2d 881, 884–85 (5th Cir. 1976). Phillips challenges only the sufficiency of the evidence supporting the determination that he took a substantial step toward commission of attempted Hobbs Act robbery. He points to his lack of participation and communication on the day of the attempted murder-robbery to support his argument. On the other hand, the Government argues that there was abundant evidence that Phillips took substantial steps toward committing the offense. The Government argues that, taken together, the evidence it presented about Phillips’s participation in the conspiracy to commit the attempted murder-robbery conclusively corroborates the firmness of Phillips’s criminal intent. Howard, 766 F.3d at 419. We agree that sufficient evidence supports the jury’s determination that Phillips took the substantial step necessary to convict him of Count Three. Phillips’s lack of participation on the day of the attempted murder-robbery does not negate the substantial evidence presented that Phillips intended and took substantial steps toward committing the offense, including Phillips’s recruitment of Duncan-Bush to the scheme, his delivery of Duncan-Bush’s burner phone, and his compliance with Batiste’s orders to drive Duncan-Bush to a nearby hotel to review plans ahead of the bank robbery. Further, Phillips was not required to participate in the attempted murder-robbery’s final acts in order to take a “substantial step.” Howard, 766 F.3d at 419. Taking the evidence in the light most favorable to the verdict, Phillips at the very least was integral in recruiting Duncan-Bush and facilitating and directing his participation in the attempted Amegy Bank robbery. On this record, “a rational trier of fact could have found the elements of the crime beyond a reasonable doubt.” Dailey, 868 F.3d at 327. Count Four Aiding and abetting a § 924(c)(1)(A)(iii) violation has four base elements: (1) that the offense of [discharge of a firearm in relation to a crime of violence (here attempted Hobbs Act robbery)] was committed by some person; (2) that the defendant associated with the criminal venture; (3) that the defendant purposefully participated in the criminal venture; and (4) that the defendant sought by action to make that venture successful. 5th Cir. Pattern Jury Instructions (Criminal) § 2.04; see also United States v. Bowens, 907 F.3d 347, 351–52 (5th Cir. 2018) (approvingly citing these instructions as elements), cert. denied, 139 S. Ct. 1299 (2019). The Supreme Court’s decision in Rosemond v. United States added a fifth element unique to an aiding-and-abetting-§ 924(c) offense: the defendant must have “advance knowledge that a confederate would use or carry a gun during the crime’s commission.” 572 U.S. 65, 67 (2014); see also 5th Cir. Pattern Jury Instructions (Criminal) § 2.04 (note). Advance knowledge “means knowledge at a time the accomplice can do something with it—most notably, opt to walk away.” Rosemond, 572 U.S. at 78. Phillips argues that the Government failed to prove the final element: that he had advance knowledge that a gun would be used in the commission of the crime. The Government argues that it defies common sense to believe that Phillips could not reasonably foresee that a firearm would be used in the commission of the offense. United States v. Saunders, 605 F. App’x 285, 289 (5th Cir. 2015) (unpublished). Moreover, the Government argues that the jury had enough evidence to find that element met beyond a reasonable doubt. We agree with the Government that, viewing, as we must, all evidence in the light most favorable to the jury’s verdict, the evidence was sufficient to support Phillips’s Count Four conviction. See, e.g., Dailey, 868 F.3d at 327. Christopher Braziel, Phillips’s cellmate and a witness for the Government at trial, testified that Phillips had told him that the group’s plan was to “kill the cash carrier and grab the money.” Duncan-Bush also testified to Phillips’s role in recruiting him to participate in the crime, and this testimony was supported by the wiretapped calls between Phillips and Batiste. Additionally, the wiretapped calls reveal that Batiste called Phillips and mentioned using an AR-15 semi-automatic rifle, as well as telling Phillips that he had had the gun’s ballistics modified in case law enforcement recovered ballistic evidence from the shooting. On “highly deferential” sufficiency review, United States v. Zamora-Salazar, 860 F.3d 826, 831 (5th Cir. 2017), there was sufficient evidence to support this verdict. Is Attempted Hobbs Act Robbery a Crime of Violence? All four defendants argue that attempted Hobbs Act robbery is not a crime of violence under 18 U.S.C. § 924(c)’s elements clause and thus cannot support their convictions for Counts Two and Four. The Government argues that this argument is foreclosed by this court’s precedents. This court reviews the legal question of whether a predicate offense qualifies as a crime of violence de novo. See, e.g., United States v. Smith, 957 F.3d 590, 592 (5th Cir. 2020), cert. denied, 141 S. Ct. 828 (2020). In United States v. Davis, the Supreme Court held that the residual clause of § 924(c)(3)(B) was unconstitutionally vague. 139 S. Ct. 2319, 2366 (2019). But a defendant’s § 924(c) “convictions can still be sustained if the predicate offenses. . . can be defined as a [crime of violence] under the elements clause contained in § 924(c)(3)(A).” Smith, 957 F.3d at 592–93. Additionally, “[w]hen a substantive offense would be a crime of violence” under the elements clause, “an attempt to commit that offense is also a crime of violence.” Id. at 595 (internal quotation marks omitted). Our precedents establish that Hobbs Act robbery is a crime of violence under the elements clause. See Bowens, 907 F.3d at 353–54 (“As the government correctly notes, binding circuit precedent forecloses Bowens’s claim that Hobbs Act robbery is not a [crime of violence] predicate[.]“). The Defendants point out that this circuit has held that conspiracies to commit a felony are not categorically crimes of violence because, considering the least culpable conduct by which a conspiracy to commit a felony can be committed, conspiracy may be committed in a manner that does not include the use, attempted use, or threatened use of physical force. United States v. Reece, 938 F.3d 630, 636 (5th Cir. 2019). But attempt is fundamentally different from conspiracy. We have held that “a predicate attempt offense that includes the specific intention to commit a [crime of violence] and a substantial step to bring about or accomplish that [crime of violence], is in and of itself a [crime of violence] under the elements clause.” Smith, 957 F.3d at 595–56. Thus, this argument is foreclosed. Sentencing Enhancement U.S.S.G. § 2B3.1(c)(1) Finally, Phillips, Polk, and Scott contend that the district court erred in applying sentencing enhancement U.S.S.G. § 2B3.1(c)(1) to their Count Three convictions based on the killing of Batiste by police during the attempted Amegy Bank ATM robbery. This enhancement, in pertinent part, instructs district courts to apply § 2A1.1, the first-degree-murder Guideline, “[i]f a victim was killed under circumstances that would constitute murder under 18 U.S.C. § 1111[.]” Phillips renews his argument on appeal that the district court erroneously applied this sentencing enhancement because Batiste was not a “ victim” under the meaning of the Guidelines; that is, that a coconspirator who is killed during the commission of a crime does not constitute a “victim” for the purpose of applying this enhancement. Additionally, Polk and Scott, along with Phillips, raise the new theory that this enhancement was erroneously applied because Batiste’s killing by law enforcement was not a killing “under circumstances that would constitute murder under 18 U.S.C. § 1111[.]“ When an issue is preserved, this court reviews the district court’s interpretation of the Guidelines de novo and its underlying factual findings for clear error. See, e.g., United States v. Narez-Garcia, 819 F.3d 146, 149 (5th Cir. 2016). But when “the basis for the defendant’s objection during trial is different from the theory [he or] she raises on appeal[,]” this court’s review is for plain error. United States v. Sanders, 952 F.3d 263, 282 (5th Cir. 2020) (cleaned up). In the Guidelines context, an objection in the district court to an enhancement on one ground does not preserve for appeal alternative arguments against that enhancement. Narez-Garcia, 819 F.3d at 149. Thus, we review Phillips’s preserved challenge de novo and Polk and Scott’s newly raised claim for plain error. Polk, Scott, and Phillips raise an unpreserved challenge to the application of sentencing enhancement U.S.S.G. § 2B3.1(c)(1), arguing that Batiste’s killing by law enforcement was not a killing “under circumstances that would constitute murder” under the Guidelines’ definition. They argue that this court should impose felony-murder liability under a theory of agency liability, rather than a proximate cause theory. U.S.S.G. § 2B3.1(c)(1). Under agency theory, the felony murder doctrine does not allow the killing of a coconspirator by police to be imputed to his fellow conspirator because the police do not act as agents of the conspiracy; however, the proximate cause theory does allow this imputation, as the commission or attempted commission of the underlying crime is still the proximate cause of the killing by police. See, e.g., Moore v. Wyrick, 766 F.2d 1253, 1255–56 (8th Cir. 1985) (defining and contrasting these two theories of felony murder liability). This new challenge cannot succeed on plain error review. As the Government points out, the Defendants cite no binding caselaw which adopts either the agency theory or the proximate cause theory of felony murder in this context. Thus, any potential error is not plain. See United States v. Gonzalez, 792 F.3d 534, 538 (5th Cir. 2015) (holding that a “lack of binding authority is often dispositive in the plain-error context”); see also United States v. McNabb, 958 F.3d 338, 341 (5th Cir. 2020) (“By definition, a close call cannot be the obvious or plain error a defendant needs to show when asserting an error he did not give the district court a chance to fix.”). Whether Phillips’s challenge to the classification of Batiste as a “victim” can prevail on de novo review is a more complicated question. As Phillips points out, we held in United States v. Geeslin, 447 F.3d 408 (5th Cir. 2006) that, for the purposes of a different provision of the Guidelines, § 2B1.1(b)(1), a participant in a crime whose actions were “not entirely voluntary” could be considered a victim, calling this a “rare circumstance[.]” Id. at 411. Phillips’s point is well taken that if it is a rare circumstance in which a coconspirator can be considered a victim for sentence enhancement purposes, it would seem strange to deem Batiste, the mastermind of this robbery scheme, a victim. Nonetheless, we need not decide this issue because the record demonstrates that any potential error in applying the sentencing enhancement was harmless. See, e.g., United States v. Groce, 784 F.3d 291, 296 (5th Cir. 2015) (declining to resolve a “not entirely clear” Guidelines issue based on harmless error). “A procedural error” in applying the Guidelines “is harmless if the error did not affect the district court’s choice of sentence.” United States v. Halverson, 897 F.3d 645, 652 (5th Cir. 2018). There are “at least two methods for the Government to show that the district court would have imposed the same sentence.” United States v. Vega-Garcia, 893 F.3d 326, 327 (5th Cir.) (per curiam), cert. denied, 139 S. Ct. 441 (2018). The first requires the Government to demonstrate “that the district court considered both ranges (the one now found incorrect and the one now deemed correct) and explained that it would give the same sentence either way.” Id. (internal quotation marks omitted). The second requires “the Government to convincingly demonstrate both (1) that the district court would have imposed the same sentence had it not made the error, and (2) that it would have done so for the same reasons it gave at the prior sentencing.” Id. (internal quotation marks omitted). Whichever the method, “[a]lthough clarity of intent must be expressed, such statements do not require magic words.” United States v. Shepherd, 848 F.3d 425, 427 (5th Cir. 2017). Here, the district court made explicit that it was aware of the objection to this sentencing enhancement and the differences in the Guidelines ranges for the Defendants if it did not apply the enhancement. The court stated explicitly that it would have imposed the same sentence under the 18 U.S.C. § 3553(a) factors even if the murder cross-reference did not apply. Thus, the Defendants cannot show that any potential error affected their substantial rights. We therefore affirm the district court’s application of the sentencing enhancement. Conclusion For the foregoing reasons, we AFFIRM the Defendants’ convictions and sentences.