Jury Exposure to Unadmitted Exhibits (Including a Codefendant Statement) Is Not Structural Error; Courts Must Apply Harmless-Error Review

1. Introduction

Case: United States v. Erik Charles Maund; Bryon Brockway; Adam Carey, No. 24-5932 (6th Cir. Feb. 23, 2026).
Parties: The United States (Plaintiff-Appellant) appealed an order granting new trials to Maund, Brockway, and Carey (Defendants-Appellees).
Background: A jury convicted the defendants of offenses arising from a Nashville murder-for-hire plot culminating in the killings of Holly Williams and William Lanway. After the verdict, the district court discovered that the jury, during deliberations, received multiple unadmitted exhibits and did not receive some admitted exhibits. The most contested items were Carey Exhibits 3 and 4—an unredacted recording and transcript containing Brockway’s statement: “Adam [Carey] didn’t know any of this shit.”
Key issues on appeal:

  • Whether the jury’s exposure to unadmitted exhibits (with no curative instruction because the mistake was discovered post-verdict) constituted structural error requiring automatic reversal.
  • If not structural, whether harmless-error review applies and whether the error was harmless as to each defendant.
  • How to conceptualize the error given its overlap with Remmer (extraneous influence) and Bruton (codefendant statement) doctrines.

2. Summary of the Opinion

The Sixth Circuit (Judge Ritz) reversed the district court’s grant of new trials. The court held:

  • The error—jury exposure to unadmitted exhibits—was not structural; it does not “defy analysis by harmless-error standards.”
  • Harmless-error review applies, even though (i) the jury received the information with the “imprimatur of the Court,” and (ii) the error was discovered only after the verdict, eliminating the possibility of a contemporaneous curative instruction.
  • Applying the most demanding harmlessness standard (“harmless beyond a reasonable doubt”), the error was harmless as to Carey, Maund, and Brockway.

The case was remanded for further proceedings consistent with reinstating the verdicts.

3. Analysis

3.1. Precedents Cited (and How They Shaped the Holding)

A. Structural error framework and the “too hard to measure” rationale

  • Arizona v. Fulminante, 499 U.S. 279 (1991): Supplied the core distinction between “trial errors” (typically harmless-error reviewable) and “structural defects” (affecting the trial framework). The court used Fulminante to classify the jury-exhibit mistake as a process error, not a framework defect.
  • Weaver v. Massachusetts, 582 U.S. 286 (2017): Provided the modern taxonomy for structural error (three independent rationales), including the “effects too hard to measure” category invoked by the district court. The Sixth Circuit treated Weaver as narrowing, not expanding, structural error and rejected defendants’ invitation to recognize a new structural-error category for this kind of exhibit mix-up.
  • Neder v. United States, 527 U.S. 1 (1999): Emphasized the Supreme Court’s “traditional categorical approach” to structural errors and warned against creating new structural categories for factually unique mistakes. The panel relied on Neder to resist converting exhibit-delivery mistakes into automatic-reversal errors.
  • Sullivan v. Louisiana, 508 U.S. 275 (1993) and Chapman v. California, 386 U.S. 18 (1967): Sullivan underscored that not all constitutional errors require reversal; Chapman provided the benchmark harmlessness test (“harmless beyond a reasonable doubt”) used as the most stringent standard in the panel’s analysis.
  • Sixth Circuit illustrations of what does qualify as structural: United States v. Liggins, 76 F.4th 500 (6th Cir. 2023) (recusal), Cunningham v. Shoop, 23 F.4th 636 (6th Cir. 2022) (biased juror referenced), Hendrickson, 822 F.3d 812 (6th Cir. 2016) (self-representation), Doan v. Carter, 548 F.3d 449 (6th Cir. 2008) (burden-of-proof-altering instructions), Batson v. Kentucky, 476 U.S. 79 (1986) (discriminatory jury selection), Greer v. United States, 593 U.S. 503 (2021) and United States v. Davila, 569 U.S. 597 (2013) (public-trial principles). These examples served as contrast: the panel portrayed the exhibit mistake as far less “framework” altering.

B. Remmer, extraneous influence, and post-verdict inquiry

  • Remmer v. United States, 347 U.S. 227 (1954): Anchored the district court’s choice to hold a “Remmer hearing” to assess juror exposure and prejudice. The Sixth Circuit accepted that this scenario can be analyzed under Remmer because the key question is exposure to an influence external to the jury.
  • In re Sittenfeld, 49 F.4th 1061 (6th Cir. 2022): Defined the Sixth Circuit’s Remmer-hearing trigger (a “colorable claim” of obvious/likely adverse effect) and reiterated the circuit’s rule that the defendant bears the burden of “actual prejudice” in Remmer settings—though the panel ultimately sidestepped resolving which burden standard should apply here.
  • United States v. Walker, 1 F.3d 423 (6th Cir. 1993): Supported the panel’s view that the source/form of the communication is less important than its potential to affect impartiality. This undercut defendants’ argument that the court’s own mistake is categorically different from classic Remmer scenarios.
  • Warger v. Shauers, 574 U.S. 40 (2014), plus Rule 606(b) applications: United States v. Bailey, No. 19-2280, 2022 WL 2444930 (6th Cir. July 5, 2022), United States v. Brooks, 987 F.3d 593 (6th Cir. 2021), Tanner v. United States, 483 U.S. 107 (1987), Parker v. Gladden, 385 U.S. 363 (1966). These authorities framed the “external vs. internal influence” line: unadmitted exhibits are “external” to the jury even if the court inadvertently supplied them.
  • Comparative Remmer-like modern examples: Ewing v. Horton, 914 F.3d 1027 (6th Cir. 2019) (internet research), United States v. Kechego, 91 F.4th 845 (6th Cir. 2024) (phone use during deliberations). These cases reinforced that courts routinely measure prejudice from extraneous influences rather than treating them as immeasurable.
  • United States v. Lanier, 870 F.3d 546 (6th Cir. 2017) and Cunningham v. Shoop, 23 F.4th 636 (6th Cir. 2022): Cited for requiring a Remmer hearing to measure impact. The panel used these cases to show that “measurement” is an ordinary judicial task, not a reason to declare structural error.

C. Bruton and confrontation concerns

  • Bruton v. United States, 91 U.S. 123 (1968): The government had redacted Brockway’s “Adam didn’t know” statement from its admitted exhibits specifically to avoid a Bruton problem (a nontestifying codefendant’s statement naming another defendant). The accidental submission of the unredacted statement during deliberations reintroduced the Bruton-adjacent risk.
  • Brown v. United States, 411 U.S. 223 (1973) and Harrington v. California, 395 U.S. 250 (1969): Established that Bruton errors are not structural and can be harmless. The panel relied on these to reinforce the broader theme: even serious confrontation errors are measured for harmlessness.
  • Samia v. United States, 599 U.S. 635 (2023) (quoting Richardson v. Marsh, 481 U.S. 200 (1987)): Provided the governing Bruton formulation and limitations, which informed the court’s explanation of why harmless-error review is appropriate even for codefendant-statement problems.
  • United States v. Alkufi, 636 F. App’x 323 (6th Cir. 2016): Offered the principle that statements implicating a defendant only in light of other evidence may not violate the Confrontation Clause—relevant context for assessing the statement’s practical prejudicial force.

D. Prior “unadmitted material to the jury” cases and the Ninth Circuit outlier

  • Nevers v. Killinger, 169 F.3d 352 (6th Cir. 1999), abrogated on other grounds by Harris v. Stovall, 212 F.3d 940 (6th Cir. 2000): Central comparator. The Sixth Circuit previously approved post-verdict harmless-error analysis where jurors received unadmitted negative news reports/rumors (no curative instruction). The panel used Nevers to demonstrate that even inflammatory extraneous material is not automatically structural.
  • United States v. Blanton, 719 F.2d 815 (6th Cir. 1983): Quoted to caution against elevating uniqueness into structural error—every trial is unique to some degree.
  • United States v. Noushfar, 78 F.3d 1442 (9th Cir. 1996), amended, 140 F.3d 1244 (9th Cir. 1998), and Eslaminia v. White, 136 F.3d 1234 (9th Cir. 1998): The defendants relied on Noushfar (structural error when unplayed tapes sent to jury). The Sixth Circuit distinguished it (inadvertence here; no “abdication of judicial control”) and emphasized that the Ninth Circuit itself limited Noushfar to extreme facts.

E. Harmless-error mechanics and standards

  • United States v. Hendrickson, 822 F.3d 812 (6th Cir. 2016) and United States v. Taylor, 127 F.4th 1008 (6th Cir. 2025): Supported the appellate court’s authority to conduct harmless-error analysis “in the first instance” when the district court applied the wrong legal standard.
  • United States v. Macias, 387 F.3d 509 (6th Cir. 2004) and Delaware v. Van Arsdall, 475 U.S. 673 (1986): Provided the “overwhelming evidence” articulation and the “reasonable possibility” benchmark for whether an error contributed to a conviction.
  • United States v. Henderson, 626 F.3d 326 (6th Cir. 2010): Confirmed Chapman’s beyond-a-reasonable-doubt harmlessness standard in the Bruton context.
  • United States v. Corrado, 227 F.3d 528 (6th Cir. 2000), and United States v. Lanier, 988 F.3d 284 (6th Cir. 2021): Highlighted the Sixth Circuit’s distinctive Remmer-burden approach (defendant must prove actual prejudice). The panel flagged the tension but avoided choosing because the government prevailed even under the most stringent standard.

F. Rule 33 and standards of review

  • United States v. Robinson, 99 F.4th 344 (6th Cir. 2024): Reiterated that Rule 33 relief requires “substantial legal error” and that “interest of justice” is the touchstone.
  • United States v. Willis, 257 F.3d 636 (6th Cir. 2001) and United States v. Munoz, 605 F.3d 359 (6th Cir. 2010): Clarified review standards—de novo for the legal standard (e.g., structural vs. harmless), abuse of discretion for the ultimate new-trial decision.
  • Washington v. Recuenco, 548 U.S. 212 (2006): Treated structural error as a legal question and reinforced the presumption that most errors are harmlessness-reviewable.
  • United States v. Miller, 531 F.3d 340 (6th Cir. 2008) and United States v. Simmons, 797 F.3d 409 (6th Cir. 2015): Reaffirmed that most errors are subject to harmless-error analysis and that structural errors are rare automatic-reversal exceptions.

3.2. Legal Reasoning

A. Why the error was not “structural”

The district court labeled the mistake structural because its effects were “too hard to measure.” The Sixth Circuit rejected that reasoning for three main reasons:

  1. Doctrinal fit: The court placed the error outside the recognized “too hard to measure” structural categories exemplified by denial of counsel of choice and denial of a proper jury verdict (citing the “primary examples” and referencing decisions such as McCoy v. Louisiana and Sullivan).
  2. Administrability and experience: Courts routinely measure prejudice from extraneous influences (Remmer) and codefendant statements (Bruton). The panel emphasized that if courts can evaluate prejudice from unadmitted news reports and rumors of racial bias (Nevers v. Killinger), they can evaluate a “few unredacted sentences.”
  3. Categorical restraint: The court refused to create a new structural-error category based on a factually unique mistake, citing the Supreme Court’s preference for a categorical approach (Neder) and warning that uniqueness is not the test (Blanton).

B. Treating the mistake as Remmer/Bruton-adjacent still leads to harmless-error review

The panel acknowledged that the situation “implicates both Remmer and Bruton” because:

  • The jury encountered unadmitted material (an “external” influence relative to the jury’s deliberative record), consistent with Remmer-type concerns.
  • The material included a nontestifying codefendant’s statement that named another defendant (“Adam”), which is the kind of evidentiary risk Bruton doctrine addresses.

But crucially, the court held it did not need to pin the error to one doctrinal box because both Remmer and Bruton errors are subject to harmless-error review.

C. Harmlessness “in the first instance” and the burden-of-proof tension

Because the district court stopped at structural error and did not conduct a harmlessness analysis, the Sixth Circuit performed the analysis itself (citing Hendrickson and Taylor). The court also highlighted (without resolving) an intra-circuit tension:

  • Bruton errors: government must show harmlessness beyond a reasonable doubt (United States v. Henderson).
  • Remmer errors (Sixth Circuit rule): defendant must prove actual prejudice (In re Sittenfeld; United States v. Corrado), a position described as unique among circuits (United States v. Lanier (2021)).

The panel avoided choosing by assuming the “most stringent” standard—harmless beyond a reasonable doubt—and still finding harmlessness.

D. Defendant-by-defendant harmlessness findings

  • Carey: The improper exhibits were exculpatory to Carey (“Adam didn’t know…”), and the jury did not receive the rebuttal statement (Peled’s proffered testimony that Carey said “nobody cares about them”). Given other strong evidence (recordings, witness testimony, surveillance), the panel concluded the error did not contribute to Carey’s conviction and may have helped him.
  • Maund: The “Adam didn’t know…” statement did not directly bear on Maund’s guilt for murder-for-hire conspiracy. The government’s evidence—Maund’s recorded call, Peled’s direct incriminating testimony about Maund, coworker testimony, and financial transfers—was substantial. Any effect on Maund’s theory that Carey acted alone was “tenuous.”
  • Brockway: The court recognized Brockway had the “strongest claim” of potential prejudice because the statement could undermine a defense that Carey acted alone. Still, given overwhelming evidence (Brockway’s recorded discussion, Peled’s testimony, travel timing, rental car matching surveillance), the panel deemed the statement “insignificant by comparison.”

E. The Remmer-hearing “binder color” mix-up

The court also rejected Maund’s claim that the hearing was constitutionally deficient under United States v. Lanier, 988 F.3d 284 (6th Cir. 2021). Although the parties were misled about exhibit-binder consolidation post-verdict, the panel concluded this did not “shackle” the defense investigation into the external influence—particularly where jurors often could not recall binder colors.

3.3. Impact

  • Limits expansion of structural error: The decision resists a broader structural-error doctrine for deliberation-phase exhibit mishaps, even when discovered only after verdict and even when the materials appear to be court-sanctioned exhibits.
  • Normalizes post-verdict prejudice measurement: The opinion reinforces that courts can (and should) evaluate prejudice from jury exposure to unadmitted items through hearings and harmless-error analysis, rather than presuming automatic reversal.
  • Practical guidance for litigants: Defendants should be prepared to articulate concrete mechanisms of prejudice (not just procedural irregularity), especially where the extraneous material is brief or ambiguous in effect.
  • Highlights unresolved burden tension: By expressly noting the Sixth Circuit’s unusual Remmer burden allocation while avoiding a definitive ruling, the opinion preserves a live issue for future cases where harmlessness is closer.
  • Co-defendant dynamics matter: The case illustrates how an “extraneous” codefendant statement may be simultaneously exculpatory to one defendant and prejudicial to another—and that harmlessness can be defendant-specific.

4. Complex Concepts Simplified

  • Structural error: A rare type of error so fundamental that appellate courts reverse automatically without asking whether it changed the outcome (e.g., denial of counsel, biased judge/jury, altered burden of proof).
  • Harmless error: An error that occurred, but the conviction stands if the appellate court is satisfied it did not affect the verdict (often phrased as “harmless beyond a reasonable doubt” for constitutional errors).
  • Remmer hearing: A procedure for investigating whether jurors were exposed to external information or contacts and whether that exposure prejudiced the verdict.
  • Bruton problem: In a joint trial, admitting a nontestifying codefendant’s confession that directly incriminates another defendant can violate the other defendant’s confrontation rights.
  • “Extraneous” vs. “internal” jury influences (Rule 606(b)): Jurors generally cannot testify about internal deliberations (pressure, reasoning), but can testify about outside information reaching the jury (internet research, third-party communications, unadmitted exhibits).
  • Rule 33 (new trial): Allows a district court to grant a new trial “if the interest of justice so requires,” but appellate review demands using the correct legal framework (structural vs. harmless).

5. Conclusion

United States v. Maund establishes a clear Sixth Circuit rule that inadvertent submission of unadmitted exhibits to the deliberating jury—even including an unredacted codefendant statement and even without a curative instruction because the error is discovered post-verdict—is not structural error. Courts must instead perform a harmless-error assessment. Applying that framework, the Sixth Circuit found the exhibit-delivery mistake harmless as to all three defendants and reversed the district court’s new-trial order. The decision narrows the path to automatic reversal for deliberation-stage irregularities and reinforces outcome-focused prejudice analysis as the governing approach.