Background
A jury convicted Adolfo Miranda in 2015 on six counts of aggravated sexual abuse of a child and three counts of rape of a child for abuse he perpetrated against his stepdaughter over a period of several years beginning when she was around nine years old. The victim testified at trial that Miranda had abused her repeatedly across three geographic settings—the family home, a Provo mini mall where Miranda worked, and an Orem arcade—and that he had raped her five or six times, primarily in the basement bedroom of the family home. Miranda took the stand and denied the allegations, suggesting the victim had fabricated them at her mother’s instigation during contentious divorce proceedings. The jury returned guilty verdicts on all nine counts in less than 90 minutes of deliberation. Miranda’s direct appeal was unsuccessful. State v. Miranda (Miranda I), 2017 UT App 203, 407 P.3d 1033.
In post-conviction proceedings, Miranda argued that both his trial counsel and his direct-appeal counsel were ineffective for failing to request a specific unanimity instruction—an instruction that would have required the jury to agree on the same specific act to support each count. Under Utah’s constitution, a jury verdict must be unanimous “as to a specific crime” and “on all elements of a criminal charge.” State v. Hummel, 2017 UT 19, ¶ 28–29. In multiple-acts cases where the evidence details more criminal acts than the counts charged, instructions must either link each charge to a specific act or require jurors to agree unanimously on the same underlying act. The State conceded that both trial and appellate counsel performed deficiently by not addressing the unanimity requirement, leaving only the prejudice prong of Strickland v. Washington, 466 U.S. 668 (1984), at issue.
The post-conviction court vacated Miranda’s convictions and ordered a new trial, relying on State v. Chadwick, 2024 UT 34, 554 P.3d 1098, for the proposition that multiple-acts cases involving identical, unlinked charges “inherently undermine confidence in the unanimity of a verdict.” The court also faulted the prosecutor’s closing argument for telling jurors that the six sexual abuse counts were “a representative example” of many incidents, which it said encouraged jurors to “gloss over the specific facts required to unanimously establish each element of each separate crime.” The State appealed.
The Court’s Holding
The Utah Court of Appeals (Judge Orme, joined by Judges Mortensen and Oliver) reversed the post-conviction court and reinstated all nine convictions. The court’s analysis proceeded in two parts.
First, the post-conviction court misread Chadwick. That case addressed the prejudice standard for preserved unanimity errors—errors raised at trial and thus subject to constitutional harmless-error review, which carries a rebuttable presumption of prejudice. Chadwick, 2024 UT 34, ¶ 57. But Miranda’s claim arose from an unpreserved error brought through an ineffective-assistance vehicle. Ineffective-assistance claims, including unpreserved unanimity challenges, are governed by Strickland’s prejudice standard, which requires the defendant to show “a reasonable probability that, but for counsel’s unprofessional errors, the result of the proceeding would have been different.” Strickland, 466 U.S. at 694. The defendant bears that burden; there is no presumption. State v. Baugh, 2024 UT 33, ¶ 42.
Second, applying Strickland, the court concluded Miranda had not demonstrated prejudice. On the three rape-of-a-child counts, the prosecutor’s closing argument identified the evidence of rape as occurring in two distinct settings—the basement bedroom (multiple times) and a friend’s house (once). Given that the jury unanimously convicted on all three counts, it necessarily agreed Miranda raped the victim in the basement at least twice, and then the third time as well. No basis existed for concluding any juror believed Miranda raped the victim only twice in the basement but rejected the others. On the six aggravated sexual abuse counts, the prosecutor specifically linked each pair of counts to a geographic location: home (counts 1 and 2), mini mall (counts 3 and 4), arcade (counts 5 and 6). While the prosecutor did not identify the specific type of touching at each location, this “geographical link was sufficient to alleviate any potential prejudice.” The court observed that “on this record we see it as highly unlikely that the jury would not have agreed on the touching that occurred.” The trial court’s characterization of the victim’s testimony as “impressive and resolute” reinforced this confidence.
Key Takeaways
- Utah post-conviction courts may not apply Chadwick’s rebuttable presumption of prejudice to ineffective-assistance claims involving the absence of a specific unanimity instruction. The presumption applies only to preserved errors; Strickland’s defendant-bears-the-burden standard governs ineffective-assistance claims.
- A prosecutor’s closing argument that geographically links each count to a distinct location or setting can substantially cure a unanimity risk—even without specifying the precise act within that location—where the victim’s testimony is detailed and credible.
- Trial court credibility findings made during the original proceeding remain available to inform the prejudice analysis at the post-conviction stage.
- Defense counsel in multiple-acts cases should request a specific unanimity instruction that links each count to a specific act, and should build a record of any strategic reason not to do so, to preserve both the issue on direct appeal and any later ineffective-assistance claim.
Why It Matters
Miranda establishes a clear methodological rule for Utah post-conviction courts evaluating unanimity-based ineffective-assistance claims: the Chadwick presumption is reserved for preserved errors. This distinction is significant because many post-conviction petitions in child sexual abuse cases raise unanimity challenges on ineffective-assistance grounds. Without a presumption to rely on, petitioners must affirmatively demonstrate a reasonable probability that the jury would have acquitted on at least one count had it been properly instructed—a difficult showing when a victim’s testimony is credible, detailed, and covered multiple incidents at multiple locations.
For prosecutors, the decision provides a practical guide. In multiple-acts cases where a specific unanimity instruction is not requested, closing arguments that clearly segregate the evidence by location or factual circumstance can substantially reduce the risk that a future unanimity challenge will succeed on prejudice grounds. The decision also serves as a reminder that the jury’s partial acquittal—or, as here, its rapid and comprehensive guilty verdict after hearing detailed testimony—is a significant data point in the prejudice analysis.