When Delayed Disclosure Is the Core Defense, Removing It from Credibility Review Is Reversible Error

Case: State v. Jose T. (Concurrence & Dissent)

Court: Supreme Court of Connecticut

Date: September 29, 2026

Author: Justice McDonald, joined by Justices D’Auria and Ecker

Introduction

The case arose from allegations that Jose T. sexually assaulted the complainant when she was approximately five or six years old. The alleged conduct occurred in 2008 or 2009, but the complainant did not disclose it until 2018. The prosecution therefore depended largely on the jury’s assessment of the complainant’s credibility more than a decade after the alleged events.

The complainant offered several explanations for the delay. She testified that a school assignment triggered flashbacks and brought back buried memories. At other points, she said that she had never forgotten the abuse but remained silent because of fear, threats, concern for her sisters, negative reactions from schoolmates, and gifts or “bribes” from the defendant.

The defense argued that these explanations were inconsistent, that the complainant had significant memory lapses, that the police failed to investigate the alleged school-assignment trigger adequately, and that the accusations were fabricated after a financial dispute concerning expenses for the complainant’s quinceañera.

The central issue in Justice McDonald’s opinion was whether the trial court’s erroneous instruction—telling jurors that they could not consider the “mere fact” of delayed disclosure in assessing credibility—was harmless. The majority concluded that it was. The concurrence and dissent concluded that it was not.

Summary of the Opinion

Justice McDonald agreed with the majority that the challenged delayed-disclosure instruction misstated Connecticut law. He nevertheless rejected the majority’s harmless-error analysis for three principal reasons:

  1. Delayed disclosure was central to both parties’ theories. The prosecution repeatedly relied on delayed-disclosure evidence and expert testimony to bolster the complainant’s credibility, while the defense used the shifting explanations for delay to argue fabrication and unreliability.
  2. The state’s case was not particularly strong. It rested predominantly on the complainant’s testimony, lacked physical evidence and direct corroboration, and involved inconsistencies that could reasonably have produced doubt.
  3. The instruction materially restricted the jury’s credibility assessment. General instructions about reasonable doubt, expert testimony, and witness credibility did not cure a specific direction preventing consideration of a central credibility factor.

The dissent would therefore have reversed the conviction and ordered further proceedings rather than affirming on the ground of harmless error.

Analysis

1. Historical and Legal Context

The opinion begins by contrasting modern law with the common-law “fresh complaint” and “hue and cry” doctrines. Those doctrines assumed that a genuine victim would promptly report a crime. Modern courts recognize that children may delay reporting sexual abuse for many reasons, including fear, shame, dependency, confusion, and family pressure.

Connecticut’s 2019 legislation eliminating the statute of limitations for crimes involving the sexual abuse, exploitation, or assault of a child reflects this modern understanding. Justice McDonald emphasized, however, that allowing prosecutions of very old allegations creates countervailing risks: witnesses disappear, memories fade, records are lost, and defendants may no longer be able to establish an alibi or locate exculpatory evidence.

The judicial task is therefore one of balance. Courts must reject the unsupported assumption that genuine victims always report immediately without treating delay as irrelevant in every case. Delay may be understandable and still remain part of the factual context through which credibility is assessed.

2. Delayed Disclosure as the Core of the Case

The opinion viewed delayed disclosure as more than an isolated impeachment point. It connected nearly every important dispute:

  • whether the complainant had repressed the memories or had never forgotten them;
  • whether fear, threats, family concerns, or gifts explained her silence;
  • whether later additions and changes reflected ordinary disclosure dynamics or fabrication;
  • whether the police investigation adequately tested the alleged trigger for disclosure; and
  • whether the allegations arose from sexual abuse or from a later financial conflict.

The prosecution called an expert, Monica Madigan, to explain child sexual-abuse disclosure patterns and repeatedly argued that those patterns supported the complainant’s credibility. The defense, in turn, highlighted the expert’s acknowledgment that inconsistent successive disclosures “could be” evidence of a false accusation.

Justice McDonald found a fundamental imbalance in allowing the state to use delayed-disclosure evidence affirmatively while instructing the jury not to use the delay in evaluating the defense’s competing account. Defense counsel’s statement that “[t]he delay is the whole defense” captured the significance of the disputed instruction.

3. Strength of the State’s Case

The state’s case was primarily a credibility contest. There was no physical or forensic evidence, confession, eyewitness to the alleged assaults, electronic record, or independent evidence that the defendant had sexually abused others.

The complainant’s sister offered some circumstantial support by testifying that the complainant had described playing a game in which she was the defendant’s “girlfriend.” But this evidence did not directly corroborate sexual assault. The defense also suggested that the sisters’ accounts reflected collaboration rather than corroboration.

Justice McDonald additionally noted inconsistencies in the complainant’s accounts, including changes concerning anal penetration, ejaculation, additional incidents, and her reasons for delaying disclosure. The defendant testified, denied the allegations, and supplied an alternative motive for fabrication. Although the jury was entitled to reject his testimony, its existence reinforced the conclusion that the case turned on credibility rather than overwhelming independent proof.

4. Why the General Instructions Did Not Cure the Error

The state argued that the charge as a whole correctly instructed the jury about the presumption of innocence, reasonable doubt, expert testimony, and credibility. The dissent considered those instructions insufficient. A general command to assess credibility cannot cure a specific command removing a material credibility consideration from the jury.

The state also attempted to distinguish the “mere fact” of delay from the reasons for delay. Justice McDonald found that distinction unintelligible in practice. Jurors cannot realistically evaluate the reasons for a ten-year delay without considering the delay itself, its length, the reliability of memory over time, and the surrounding circumstances.

Because courts presume that juries follow instructions, the harmlessness analysis had to assume that the jury complied with the erroneous restriction.

5. Harmless-Error Framework

The opinion reviewed Connecticut’s varying formulations of harmless-error review for nonconstitutional errors. Earlier decisions asked whether an error probably affected the result or whether it undermined confidence in the verdict. State v. Sawyer attempted to unify the inquiry by asking whether the appellate court has a “fair assurance” that the error did not substantially affect the verdict.

For Justice McDonald, the precise wording of the standard did not alter the result. The error concerned the parties’ central dispute, occurred in a comparatively weak case, and limited the jury’s ability to perform its most important function: deciding whom to believe.

Precedents Cited

Delayed Disclosure and the Modern Approach

  • State v. Troupe: Rejected the assumptions underlying the common-law fresh-complaint approach and emphasized balancing the interests of the defendant, the state, and the victim. It also supplied the framework for constancy-of-accusation evidence.
  • State v. Daniel W. E.: Discussed the evolution of delayed-disclosure doctrine and the need to accommodate competing interests. The opinion notes that it was later overruled by State v. Adam P..
  • State v. Marshall: Supported the departure from common-law assumptions that genuine sexual-assault victims necessarily report immediately.
  • State v. Sanderson: Discussed the 2019 legislation eliminating the limitations period for crimes involving the sexual abuse, exploitation, or assault of a child.
  • Patrick v. Auto-Owners Ins. Co.: Served as a contrast: delay may be inherently suspicious for offenses, such as automobile theft, that ordinarily would be reported promptly.

Statutes of Limitations and Stale Evidence

  • Neuhaus v. DeCholnoky: Explained that statutes of limitations protect against stale or fraudulent claims and promote accurate fact-finding by reducing problems caused by lost evidence, unavailable witnesses, and faded memories.

Harmless Error and Evidentiary Strength

  • State v. Grenier: Recognized that child sexual-abuse cases often rest on credibility rather than physical evidence and discussed competing formulations of the nonconstitutional harmless-error standard.
  • People v. Henderson: Reinforced the observation that sexual-abuse prosecutions often lack the conventional evidence associated with a “strong” case.
  • State v. Sawyer: Adopted the “fair assurance” formulation for nonconstitutional harmless error and observed that a child sexual-abuse case without physical evidence is not especially strong. It was overruled on other grounds by State v. DeJesus.
  • State v. Fisher: Used the formulation asking whether prejudice undermined confidence in the fairness of the verdict.
  • State v. Carlos G.: Supplied the principle that jurors are presumed to follow the trial court’s instructions, including an erroneous one.

Comparable Credibility Cases

  • State v. Fernando V.: Found reversible error when the exclusion of a defense witness deprived the jury of evidence needed to evaluate whether the complainant displayed characteristics described by the state’s delayed-disclosure expert. The case established that removing tools essential to a credibility determination cannot readily be considered harmless.
  • State v. William G.: Held an erroneous delayed-disclosure instruction harmful where credibility, inconsistent testimony, and alleged fabrication were closely intertwined. Justice McDonald considered it strongly analogous.
  • State v. Adam P.: Rejected an attempted distinction between delay itself and the reasons for delay but found the instructional error harmless on materially stronger facts. There were two largely consistent complainants, corroborating evidence, and only minimal reliance on delayed disclosure by the defense.
  • State v. Marcello E.: Recognized that close credibility contests supported by equivocal evidence are particularly vulnerable to harmful error.
  • State v. Favoccia: Was cited as another divided child sexual-abuse decision finding reversible error, illustrating the fact-intensive and frequently contested nature of harmless-error analysis.

Corroboration and Other-Misconduct Evidence

  • State v. DeJesus: Illustrated how testimony that a defendant abused other individuals in a similar manner can substantially strengthen a complainant’s credibility. No comparable independent evidence existed here.

Complex Concepts Simplified

Delayed disclosure
A report of alleged abuse made long after the event. The term does not assume that the accusation is true or false.
Fresh complaint or hue and cry
Former doctrines based on the assumption that a genuine victim would promptly report a crime. Modern sexual-assault law rejects that assumption as unreliable, especially for children.
Harmless error
A legal mistake that does not require reversal because the appellate court is sufficiently confident that it did not materially affect the verdict.
Constancy of accusation
Evidence that a complainant made substantially consistent accusations to other people, offered to support credibility.
Uncharged misconduct or propensity evidence
Evidence of similar conduct not charged in the present case, sometimes admitted under specialized rules to show a relevant pattern or tendency.
Repressed or recovered memory
The asserted loss of conscious access to a traumatic memory followed by its later return. Here, that theory conflicted with testimony that the complainant “never forgot” the alleged abuse.

Potential Impact

Although nonbinding, the opinion offers significant guidance for future delayed-disclosure cases:

  • Trial courts should avoid instructions that categorically remove delay from the jury’s credibility analysis.
  • Courts should examine whether delay is merely incidental or is intertwined with fabrication, memory, motive, and investigative-failure defenses.
  • Prosecutors who use expert testimony to explain delay may face stronger objections to instructions preventing the defense from addressing the same subject.
  • Appellate courts should assess harmlessness in light of the entire trial strategy, not by isolating the challenged instruction from closing arguments and witness examinations.
  • Errors affecting credibility are especially consequential when the prosecution lacks physical evidence or substantial independent corroboration.

Conclusion

Justice McDonald’s central position is that delayed disclosure cannot simultaneously be used to strengthen the prosecution’s case and withheld from the defense’s credibility challenge. Where delay is “the whole defense,” an instruction restricting the jury’s consideration of it deprives jurors of a central tool for deciding whom to believe.

The opinion does not revive the discredited assumption that genuine victims always report abuse promptly. Instead, it advocates a balanced rule: delayed disclosure may have innocent explanations, but its circumstances, duration, and asserted causes must remain available for fair credibility assessment. On the record presented, the dissent found no fair assurance that the erroneous instruction left the verdict unaffected.