Post-Erlinger Harmless Error in ACCA Cases: Sixth Circuit Confirms PSRs May Ground “Different Occasions” Findings and Emphasizes the Reasonable-Jury Standard

Introduction

In United States v. Marquise Graham, the Sixth Circuit affirmed an Armed Career Criminal Act (ACCA) sentence despite an “Erlinger error” — judicial factfinding that should have been submitted to a jury under the Supreme Court’s recent decision in Erlinger v. United States, 602 U.S. 821 (2024). The panel held the error was not structural and was harmless beyond a reasonable doubt, relying on undisputed facts in the Presentence Report (PSR) and other Shepard materials to conclude that a reasonable jury would have found the defendant’s prior offenses occurred on different occasions.

The case clarifies three important points in the ACCA/Erlinger landscape within the Sixth Circuit:

Erlinger errors are not structural; harmless-error review applies.
Appellate courts may rely on Shepard-approved documents and, absent specific reliability challenges, a PSR to conduct harmless-error review of the “different occasions” requirement.
Substantial temporal separation (here, an eight-month gap) and an intervening offense strongly support separate occasions; “what a reasonable jury would do” governs, not the empirics of other juries’ verdicts.

The court also rejected a post-plea Double Jeopardy challenge to the ACCA enhancement under plain-error review.

Summary of the Opinion

After officers recovered a loaded 9mm pistol from a bag the defendant discarded while fleeing a traffic stop, Marquise Graham pleaded guilty to being a felon in possession of a firearm in violation of 18 U.S.C. § 922(g)(1). The district court imposed a 192-month (16-year) sentence based on ACCA’s 15-year mandatory minimum, finding that Graham had three or more qualifying violent felonies committed on different occasions. On appeal, Graham argued:

The Erlinger error (judge, not jury, found “different occasions”) was structural and required automatic reversal.
Even if reviewed for harmless error, the court could not rely on Shepard documents or a PSR to conclude the error was harmless.
The post-plea application of an uncharged ACCA enhancement violated the Double Jeopardy Clause.

The Sixth Circuit rejected all three arguments and affirmed. It held Erlinger errors are not structural (citing circuit precedent), applied harmless-error review, and—relying on undisputed PSR facts—concluded beyond a reasonable doubt that a reasonable jury would find at least three qualifying predicate offenses occurred on different occasions. The court emphasized the strong weight of timing (an eight-month gap and an intervening offense), distinct victims and locations, and a “reasonable jury” lens. On the Double Jeopardy claim, the court applied plain-error review and found no obvious or clear violation, distinguishing earlier Sixth Circuit precedent and following its own recent decision rejecting similar arguments.

Factual and Procedural Background

In August 2021, Chattanooga police responding to a shooting pursued a suspect vehicle. Graham fled on foot, discarding a black bag in which officers found a loaded 9mm pistol and debit/credit cards bearing his name. He pleaded guilty to violating § 922(g)(1) in November 2022. The PSR identified six prior Tennessee offenses (aggravated burglaries and a robbery) with different victims and addresses in Chattanooga:

Aggravated Burglary – March 31, 2015
Robbery – May 8, 2015
Aggravated Burglary – October 11, 2015
Aggravated Burglary – October 27, 2015
Aggravated Burglary – October 27, 2015
Aggravated Burglary – November 9, 2015

Graham did not dispute at sentencing that these convictions qualify as violent felonies under ACCA or contest the PSR’s dates and locations. The district court found the offenses occurred on different occasions and imposed an ACCA sentence. Graham appealed, invoking Erlinger, which post-dated his sentencing.

Detailed Analysis

Precedents Cited and Their Influence

Erlinger v. United States, 602 U.S. 821 (2024) — Held that a jury, not a judge, must decide whether prior offenses were committed on different occasions under ACCA. Here, the district court made that finding, creating an Erlinger error. The Sixth Circuit treated this as trial error subject to harmless-error review, not a structural error mandating automatic reversal.
United States v. Campbell, 122 F.4th 624 (6th Cir. 2024) and United States v. Thomas, 142 F.4th 412 (6th Cir. 2025) — Binding Sixth Circuit authority that Erlinger errors are not structural. These cases also recognize that appellate courts may assess harmlessness using Shepard materials and, per Thomas, rely on PSR facts unless the defendant gives a specific reason to doubt their reliability.
United States v. Durham, Nos. 23-5162/5173, 2025 WL 2355998 (6th Cir. Aug. 14, 2025) — Articulates the harmless-error standard in this context: the error is harmless only if the court can conclude beyond a reasonable doubt that a jury would have found different occasions.
Wooden v. United States, 595 U.S. 360 (2022) — Provides the multi-factor framework for determining whether offenses occurred on the same “occasion,” including timing, location, character, and relationship. Wooden notes that offenses a day or more apart are “nearly always” separate occasions.
United States v. Kimbrough, 138 F.4th 473 (6th Cir. 2025) — Applies Wooden’s multi-factor analysis within the Sixth Circuit.
United States v. Ballinger, 2025 WL 2715334 (6th Cir. Sept. 16, 2025) — Clarifies that the harmless-error inquiry asks what a reasonable jury would do on the facts at hand, not what another actual jury did in a different case.
Shepard v. United States, 544 U.S. 13 (2005) and United States v. Hennessee, 932 F.3d 437 (6th Cir. 2019) — Shepard identifies the set of documents permissible to establish the nature of prior convictions (charging instrument, plea colloquy, etc.). Hennessee reflects pre-Erlinger practice where judges found “different occasions” at sentencing using Shepard documents. Post-Erlinger, judge-found “different occasions” is error, but the Sixth Circuit permits reliance on Shepard materials—and, under Thomas, an uncontested PSR—to conduct harmless-error review.
United States v. Geerken, 506 F.3d 461 (6th Cir. 2007) — A district court may rely on uncontroverted PSR facts. The panel leverages this principle in deciding there is no reason to doubt the PSR’s dates and locations, as Graham never challenged them.
United States v. Santana, 141 F.4th 847 (7th Cir. 2025) — Persuasive authority noting that a four-mile distance is not dispositive by itself; distance must be considered in context. The panel cites this to show that location alone rarely dictates the outcome.
Preservation and Plain Error: United States v. Johnson, 79 F.4th 684 (6th Cir. 2023); United States v. Ehle, 640 F.3d 689 (6th Cir. 2011); United States v. Bostic, 371 F.3d 865 (6th Cir. 2004); United States v. LeBlanc, 612 F.2d 1012 (6th Cir. 1980); United States v. Sain, No. 22-6131, 2025 WL 801366 (6th Cir. Mar. 13, 2025); United States v. Johns, 65 F.4th 891 (6th Cir. 2023); United States v. Amos, 501 F.3d 524 (6th Cir. 2007).

These authorities inform the court’s application of plain-error review to the Double Jeopardy claim and the need for specificity in preserving objections at sentencing.
United States v. Bell, 37 F.4th 1190 (6th Cir. 2022) — Distinguished. Bell’s unique facts do not control in the Erlinger/ACCA enhancement context. Thomas already rejected a similar Double Jeopardy argument, undercutting any claim of “obvious” error.

Legal Reasoning

1) Erlinger error is not structural; harmless-error review applies

The panel recognizes that, post-Erlinger, a judge’s “different occasions” finding is erroneous when not made by a jury. Relying on Campbell and Thomas, it holds the error is not structural, thereby rejecting Graham’s request for automatic reversal. The proper inquiry is whether the error was harmless—i.e., whether the court can say, beyond a reasonable doubt, that a reasonable jury would have found the prior offenses occurred on different occasions.

2) Shepard materials and an uncontested PSR may inform harmlessness

Graham argued the court could not use Shepard documents (including the PSR) for harmlessness because Erlinger requires a jury. The panel rejects this, citing Thomas: appellate courts may rely on Shepard-approved sources and PSR facts, so long as the defendant has not supplied a reason to doubt their reliability. Graham never claimed the PSR’s dates or locations were wrong; thus, the panel treats them as reliable for harmless-error analysis.

3) Applying Wooden’s multi-factor test under a “reasonable jury” lens

Taking Wooden’s factors (timing, location, character, relationship), the panel concludes beyond a reasonable doubt that a reasonable jury would find at least three offenses occurred on different occasions:

Timing: The March 31 burglary and the May 8 robbery were 38 days apart; the March burglary and the later October/November burglaries were separated by roughly eight months. Under Wooden, offenses a day or more apart are “nearly always” separate, and the lengthy gap is especially telling. The May robbery also intervenes between the March and October burglaries, reinforcing separateness.
Character and Relationship: The robbery is a different offense type from burglary. Moreover, the six predicate offenses involved six different victims in different homes—lack of repetition across the same victim or address supports separate occasions.
Location: While most addresses were within 1.6 miles, the March burglary was 3.6 miles from the cluster. Distance alone is not dispositive (cf. Santana), but, in context, the outlier distance modestly supports separateness, especially when paired with the temporal gap and intervening offense.

The panel underscores that the question is not what some other jury actually did elsewhere, but what a reasonable jury would do on these facts (Ballinger). On this record, timing and character alone are sufficient; location does not cut the other way.

4) Double Jeopardy challenge fails under plain-error review

Graham did not specifically preserve a Double Jeopardy objection below. Citing Bostic and related authorities, the court applies plain-error review and finds no “clear or obvious” error. It distinguishes Bell and follows Thomas, which already rejected similar Double Jeopardy claims in the ACCA/Erlinger posture. Absent binding precedent supporting Graham’s position, and in light of the Sixth Circuit’s own contrary authority, the claim fails.

Impact and Implications

For ACCA litigation post-Erlinger

Harmless-error framework entrenched: Within the Sixth Circuit, Erlinger errors will be reviewed for harmlessness rather than treated as structural. Appellants must be prepared to show a reasonable possibility that a jury would not find different occasions.
PSRs remain potent absent specific reliability challenges: The court reiterates that an uncontested PSR can supply the dates, locations, and offense details needed to support harmlessness. Defense counsel should object, with specificity and evidentiary support, to any PSR fact they intend to contest on appeal.
Timing is often dispositive: Following Wooden, offenses separated by a day or more are “nearly always” separate; an eight-month gap plus an intervening offense makes harmlessness especially likely. Prosecutors should marshal clear chronological and geographic data; defense counsel should develop evidence of tight temporal and locational continuity if truly present.
“Reasonable jury” lens governs: Parties should frame arguments around what a reasonable jury would conclude on the actual record, not around anomalous jury outcomes in other cases.
Intervening offenses break continuity: An intervening offense between similar crimes strongly supports finding different occasions. Defense theories of a single “criminal episode” must grapple with such breaks.

On preservation and procedural posture

Preserve, preserve, preserve: General objections about ACCA procedure will not preserve a Double Jeopardy claim. Counsel must state the “true basis” of the objection with reasonable specificity at sentencing to avoid plain-error review.
Plea practice post-Erlinger: Defendants contemplating a plea in ACCA-eligible cases should consider whether to demand a jury determination on the “different occasions” element or stipulate; prosecutors should ensure the record (including Shepard documents) is robust enough to withstand Erlinger-based appellate scrutiny.

Complex Concepts Simplified

ACCA (18 U.S.C. § 924(e)): Imposes a 15-year mandatory minimum for § 922(g) defendants who have at least three prior qualifying “violent felonies” or “serious drug offenses” committed on “occasions different from one another.”
“Different occasions”: Per Wooden, courts (and now juries post-Erlinger) assess multiple factors (timing, location, character, relationship). Significant time gaps or intervening crimes typically indicate separate occasions.
Shepard documents: Limited, reliable judicial records (e.g., charging documents, plea agreements, plea colloquies, jury instructions, judgments) used to establish the nature of prior convictions. In the Sixth Circuit’s harmless-error analysis, courts may also rely on an uncontested PSR’s factual recitals.
PSR (Presentence Report): A report prepared by probation that describes the offense, the defendant’s history, and applicable sentencing factors. Its factual statements can be taken as true if not specifically disputed.
Erlinger error: A sentencing error occurring when a judge, rather than a jury, finds that prior offenses were committed on different occasions for ACCA purposes.
Structural vs. harmless error: Structural errors require automatic reversal; harmless errors do not if the government can show beyond a reasonable doubt the verdict would have been the same absent the error.
Plain error: Appellate relief for unpreserved claims requires a clear or obvious error that affected substantial rights and seriously affected the fairness, integrity, or public reputation of judicial proceedings.
Double Jeopardy in sentencing: Typically, sentencing enhancements (like ACCA) do not constitute a separate offense. Adding or pursuing an enhancement post-plea generally does not raise Double Jeopardy concerns, absent unusual circumstances.

Assessment of the Court’s Application

The panel’s harmless-error determination is anchored by uncontested, case-specific facts: a clear eight-month hiatus between the March burglary and the October/November burglaries, an intervening robbery of a different character, and six distinct victims/addresses. Under Wooden, timing is often dispositive; the intervening offense and distinct victims/locations reinforce separateness. The court appropriately rejected appeals to anecdotal jury behavior in other cases (Ballinger), staying focused on the reasonable-jury counterfactual and the record actually before it.

On procedure, the opinion underscores an increasingly prevalent lesson in post-Erlinger litigation: defendants who neither request a jury determination nor specifically contest PSR facts risk affirmance on harmless-error grounds. Likewise, broad invocations of constitutional concerns will not preserve discrete theories such as Double Jeopardy absent clear, specific objection.

While designated “Not Recommended for Publication,” the opinion coheres with published Sixth Circuit authority (Campbell, Thomas, Kimbrough, Durham) and offers concrete, fact-patterned guidance on when timing and character resolve the “different occasions” inquiry as a matter of harmless error.

Conclusion

United States v. Graham solidifies the Sixth Circuit’s post-Erlinger approach: Erlinger errors are not structural; harmless-error review looks to what a reasonable jury would conclude on the actual record; and courts may rely on Shepard materials—and, absent specific reliability challenges, PSR facts—to find such error harmless. On the merits, substantial time gaps and intervening offenses are powerful indicators of different occasions, especially when coupled with different victims and locations. Procedurally, the opinion is a cautionary tale for preservation: generalized objections will not suffice to raise Double Jeopardy or to undermine reliance on PSR facts.

Key takeaways:

Erlinger errors are subject to harmless-error review, not automatic reversal.
Uncontested PSR facts can sustain harmlessness on the “different occasions” element.
Eight-month separation with an intervening offense is more than sufficient to establish separate occasions under a reasonable-jury standard.
Double Jeopardy challenges to post-plea ACCA enhancements face an uphill battle, particularly if not preserved with specificity.

In short, Graham reinforces a pragmatic, record-driven application of Erlinger and Wooden, providing clear guidance for courts and litigants navigating ACCA’s “different occasions” requirement in the Sixth Circuit.