The defendant, Jeffery Douglas, was convicted by a jury of first-degree and second-degree criminal sexual conduct in connection with the alleged sexual abuse of his then-three-year-old daughter, KD. Before us is whether the Court of Appeals erred in concluding that, as a result of evidentiary errors at trial and the ineffective assistance of counsel during both the pretrial and trial stages of the case, the defendant is entitled to a new trial and to the reinstatement of a plea offer he rejected. We agree with the Court of Appeals that a new trial is warranted in light of the errors by both the court and defense counsel at trial. We hold, however, that the Court of Appeals erred in concluding that the prosecution’s prior plea offer must be reinstated, as we see no reversible error in the trial court’s determination to the contrary. Accordingly, we affirm the Court of Appeals in part, reverse in part, and remand for proceedings consistent with this opinion.
I. FACTUAL AND PROCEDURAL OVERVIEW
KD is the biological daughter of the defendant and Jessica Brodie. The defendant and Brodie lived together for approximately seven years, during which time KD was born. The couple separated at the end of March 2008. Around that time, the defendant and Brodie each filed domestic violence charges against the other, which were ultimately dismissed. Upon the recommendation of Children’s Protective Services (CPS), KD went to live with the defendant in May 2008; KD was 31/z years old at the time. The defendant and KD lived with the *562defendant’s mother for approximately one month, and then lived with his current wife (then his girlfriend) from June 2008 until January 2009. At that point, KD went to live with Brodie and spent alternating weekends with the defendant. In May 2009, the defendant married his current wife and the couple announced her pregnancy shortly thereafter.
In June 2009, the instant allegations of sexual abuse surfaced: namely, that the defendant had made KD perform fellatio on him while he and KD were living with his mother approximately a year earlier, and that the defendant had made KD touch his penis on a separate, prior occasion. According to Brodie, KD spontaneously disclosed the alleged fellatio to her while the two were in the car together. As a result, Brodie moved up KD’s preexisting appointment with her therapist, who in turn contacted CPS after speaking with KD. CPS opened an investigation and, together with local police, arranged for a forensic interview of KD at Care House, a social services center committed to the prevention of child abuse. During that interview, KD discussed the alleged fellatio and touching.
The defendant was thereafter charged with one count of first-degree criminal sexual conduct (CSC-I), MCL 750.520b(l)(a), and one count of second-degree criminal sexual conduct (CSC-II), MCL 750.520c(l)(a); CPS also filed a petition to initiate child protective proceedings. KD and Brodie testified at a preliminary examination. Prior to that hearing, the prosecution discussed with defense counsel the possibility of a plea to one count of attempted CSC, which the defendant rejected. The case proceeded to trial in March 2010. Shortly beforehand, the prosecution extended a second plea offer to the defendant for one count of fourth-*563degree criminal sexual conduct (CSC-IV), MCL 750.520e, which the defendant also rejected.
At trial, the prosecution presented testimony from KD (by then five years old), Brodie, and certain individuals involved in the underlying investigation of the case: Detective Sergeant Gary Muir, who testified, in relevant part, to the content of a recorded telephone conversation between the defendant and Brodie; State Police Trooper Larry Rothman, who testified regarding two interviews he had conducted with the defendant in connection with the allegations; CPS worker Diana Fallone, who testified regarding her investigation of the allegations and decision to commence child protective proceedings; and forensic interviewer Jennifer Wheeler, who was qualified as an expert and, over the defendant’s objection, testified to the content of her interview with KD. The jury was also shown a video recording of that interview, again over the defendant’s objection.
The defendant testified in his own defense, denying any wrongdoing. The defendant also presented testimony from his mother, with whom he and KD were living at the time the fellatio was alleged to have occurred, and from his current wife. The defendant’s theory at trial was that the allegations of abuse had been fabricated by Brodie out of spite toward the defendant and his new wife, and that Brodie had coached KD accordingly.
The jury convicted the defendant as charged. As he had throughout the pretrial and trial stages of the case, the defendant maintained his innocence at sentencing. The trial court initially sentenced the defendant to concurrent prison terms of 85 to 360 months and 38 to 180 months for the CSC-I and -II convictions, respectively. After the defendant’s term of incarceration be*564gan, however, the Department of Corrections notified the court, and the court in turn notified the parties, that the defendant had not been sentenced in accordance with MCL 750.520b(2)(b), which requires a 25-year mandatory minimum sentence for his conviction of CSC-I. Neither the court, the prosecution, nor defense counsel appear to have been aware of this mandatory minimum before receiving this correspondence, and the defendant had not been informed of it at any point prior. The parties then filed competing motions: the prosecution, to modify the sentence in accordance with the mandatory minimum; the defendant, for reinstatement of the prosecution’s second pretrial plea offer, for a new trial, and for a Ginther1 hearing, claiming evidentiary errors at trial and ineffective assistance of counsel at the pretrial and trial stages.
On September 9, 2010, the trial court held a hearing on the motions, at which the defendant and his trial counsel testified; the court thereafter granted the prosecution’s motion to modify the sentence and denied the defendant’s requests for relief. The testimony received at the hearing and the court’s subsequent ruling on the motions focused predominantly on the pretrial advice the defendant had received from counsel regarding the prosecution’s plea offer and the consequences of a conviction at trial, and to what extent any errors in that advice affected the defendant’s decision to reject the offer.
The defendant appealed, and the Court of Appeals reversed, concluding that the defendant was entitled both to a new trial and to reinstatement of the prosecution’s plea offer. People v Douglas, 296 Mich App 186; 817 NW2d 640 (2012). The Court of Appeals found numerous evidentiary errors at trial, committed by *565both the court and defense counsel, that undermined the reliability of the jury’s verdict and warranted a new trial. Namely, the Court of Appeals held that the trial court erred in admitting, through Wheeler’s testimony and the video recording, KD’s out-of-court statements during the forensic interview regarding the alleged abuse. It further held that defense counsel was ineffective both for failing to object to certain inadmissible testimony from Brodie, Muir, and Fallone that bolstered KD’s credibility, and for failing to impeach KD at trial with her preliminary examination testimony. The Court of Appeals also concluded that the defendant was entitled to relief on his claim of ineffective assistance of counsel at the pretrial stage, in light of the incorrect advice counsel provided in connection with the prosecution’s plea offer. Accordingly, the Court of Appeals ordered that, upon remand, the prosecution must reoffer that plea to the defendant.2
The prosecution then sought leave to appeal in this Court, challenging both the award of a new trial to the defendant and the requirement that the prosecution’s prior plea offer be reinstated. We granted leave to appeal in order to review these issues. People v Douglas, 493 Mich 876 (2012).
II. STANDARD OF REVIEW
A trial court’s decision to admit evidence will not be disturbed absent an abuse of discretion, which occurs when the court “chooses an outcome that falls outside the range of principled outcomes.” People v Musser, 494 Mich 337, 348; 835 NW2d 319 (2013). If the court’s *566evidentiary error is nonconstitutional and preserved, then it “ ‘is presumed not to be a ground for reversal unless it affirmatively appears that, more probably than not, it was outcome determinative’ ” — i.e., that “it undermined the reliability of the verdict.” Id., quoting People v Krueger, 466 Mich 50, 54; 643 NW2d 223 (2002).
Whether the defendant received the effective assistance of counsel guaranteed him under the United States and Michigan Constitutions is a mixed question of fact and law. People v Trakhtenberg, 493 Mich 38, 47; 826 NW2d 136 (2012), citing People v Armstrong, 490 Mich 281, 289; 806 NW2d 676 (2011). This Court reviews for clear error the trial court’s findings of fact in this regard, and reviews de novo questions of constitutional law. Id.
III. THE DEFENDANT’S ENTITLEMENT TO A NEW TRIAL
We turn first to the Court of Appeals’ determination that the defendant is entitled to a new trial. We agree that such relief is warranted. This conclusion stems from errors made by both the trial court and defense counsel in the handling of evidence presented through three witnesses for the prosecution: forensic interviewer Wheeler, Detective Sergeant Muir, and CPS worker Fallone. As set forth below, the trial court erred in twice admitting the out-of-court statements made by KD to Wheeler during her forensic interview regarding the alleged fellatio; furthermore, defense counsel’s performance was constitutionally deficient in permitting Muir, Fallone, and Wheeler to offer inadmissible testimony vouching for KD’s credibility. The trial court’s error and defense counsel’s deficient performance were each sufficiently prejudicial to require a new trial.
*567A. THE EVIDENCE AT TRIAL
1. THE PROSECUTION’S CASE-IN-CHIEF
With no physical evidence of or third-party witnesses to the alleged abuse, the prosecution built its case around the credibility of KD’s in-court and out-of-court statements, and the unreliability of the defendant’s denials. The prosecution’s first witness was five-year-old KD, who testified that she sucked the defendant’s “peepee” and touched it with her hand. She initially denied that his “peepee” touched any part of her body when she sucked it, including her mouth, but later indicated that she touched it once with her hands, and once with her mouth. She also expressed uncertainty regarding what she meant by “peepee.” As to the alleged fellatio, KD indicated that it happened while she and the defendant were alone in a bedroom at the defendant’s mother’s house, that the defendant was awake and lying on a bed, and that he asked her to do it. KD testified that she told Brodie this while at Brodie’s house, and that she told Brodie the truth; she denied telling anyone but Brodie, but also indicated that she talked about it with “Jennifer” and “Tara” (whom the record indicates to be Wheeler and KD’s therapist, respectively). She affirmed that she also told Brodie that milk came out of the defendant’s “peepee” and, when asked if she told Brodie that the “milk” tasted like cherry,3 KD replied that it tasted like “peepee and regular milk.” As to the alleged touching, KD could not *568remember when it happened, but said she touched the defendant’s “peepee” with her stepsister.
The prosecution next called Brodie, who testified that in early June 2009, KD “ spontaneous [ly]” told her that “I sucked my daddy’s peepee until the milk came out, and my daddy said, oh yeah, that’s how you do it.” Contrary to KD’s testimony, Brodie indicated that this happened while she was driving in the car with KD to pick up her fiancé. When asked by Brodie, KD said this happened in the office at the defendant’s mother’s house. When asked if KD ever told her that the milk tasted like cherry, Brodie replied that KD said that at the preliminary examination but had never told her that. Brodie testified that she then moved up KD’s therapy appointment in light of the disclosure and, when CPS thereafter became involved, took KD to Care House for a forensic interview. Brodie denied that she told KD what to say; she also denied that she held any animosity toward the defendant or his new wife, or that she threatened either of them or their relationship with KD at any point prior to KD’s disclosure.
Detective Sergeant Muir then testified about his role in the investigation of these allegations. In particular, Muir testified that, after KD’s forensic interview, he asked Brodie to make a telephone call to defendant regarding the allegations. Muir recounted that Brodie told the defendant “[t]hat [KD] had said that she had sucked on her dad’s peepee and stuff came out,” and that, when the defendant responded that he did not know why KD would say that, Brodie replied, “I know my daughter don’t lie; why is she making these allegations then; was there anything that happened that, y’know, she might have seen or observed that would cause her to say this happened?” Muir further testified that Brodie and the defendant discussed an incident *569when the defendant woke up to KD touching his penis; the defendant indicated he had told Brodie this at the time, which she did not recall, and also that he had told KD “it was a bad thing to do.” Defense counsel did not object to any of this testimony.
The prosecution then presented expert testimony from Wheeler regarding KD’s forensic interview at Care House. Before Wheeler took the stand, defense counsel objected that KD’s out-of-court statements during the forensic interview were inadmissible hearsay, arguing in particular that they did not meet certain requirements of MRE 803A’s categorical hearsay exception. The trial court overruled the objection. Wheeler testified about her background and experience as a child forensic interviewer, which included “thousands” of such interviews, and was qualified as an expert in that field. After describing Care House (which she characterized as a “neutral location”) and the general protocol used for child forensic interviews, Wheeler discussed her interview with KD. She testified that KD told her that “[m]y daddy made me suck his peepee,” and that “[o]ne time we sucked it, and one time we touched it,” repeating these statements throughout her testimony and using body diagrams from the interview — including one labeled with the defendant’s name — to illustrate them. She also testified that KD told her that the alleged fellatio happened at the house of the defendant’s mother, with just her and the defendant in the room; that KD “pointed to her mouth” when asked “what did he make you suck it with”; and that KD told her the “milk” tasted like “peepee milk,” and not like cherry. Wheeler further testified that KD told her it tasted “yuk” and it went down her throat. Wheeler considered whether there had been a misunderstanding, but determined there was not because KD was “very clear” about what happened. The prosecutor then *570asked Wheeler for her opinion regarding the truthfulness of KD’s statements. After defense counsel objected, the prosecutor rephrased, asking whether Wheeler believed KD had been coached to tell her these things; without objection, Wheeler opined that KD had not. Wheeler thereafter reaffirmed that she believed KD had not been coached by Brodie, but rather “was being truthful with [her]” during the interview. Again, defense counsel did not object.
After Wheeler, the prosecution called CPS worker Diana Fallone, who testified that, in her capacity at CPS, she investigates complaints of abuse and neglect and that she performed such an investigation here. Fallone testified that, after interviewing Brodie and observing KD’s forensic interview, she filed a petition to commence child protective proceedings based on KD’s allegations. She testified that, if she thought a child were lying, she would not seek such a petition, and that she would have to substantiate that the allegations did in fact occur before seeking a petition. Fallone then testified that, based on her investigation in the instant case, she found that KD’s “allegations had been substantiated.” She further testified that, “based on the disclosures made at Care House, there was no indication that [KD] was coached or being untruthful[.]” Defense counsel did not object to this testimony.
Trooper Rothman then testified that he interviewed the defendant twice about the allegations. Rothman testified that, when he mentioned the alleged fellatio to the defendant during the first interview, defendant denied that it happened but became more nervous as the interview went on, which Rothman typically takes as a sign of untruthfulness. Rothman further testified that, during the second interview, he asked the defendant if he remembered a time when KD sucked his *571penis and the defendant responded that he did not remember that; he also asked the defendant if KD did suck his penis, and the defendant said he did not remember. The defendant told Rothman that one time when KD was approximately two, he was sleeping in the nude and woke up to find her touching his penis, which he told her not to do. The defendant also mentioned to Rothman another time when he awoke and KD and her stepsister were in his bed. In both instances the defendant stated that the children were not there when he fell asleep. Rothman was never able to substantiate the suggestion that the stepsister was involved in any touching of the defendant, and acknowledged that the stepsister, in an interview, said it did not happen.
The prosecution closed its case in chief by showing the jury the video recording of Wheeler’s forensic interview with KD. The defendant renewed his prior objection to these out-of-court statements under MRE 803A, which was again overruled. Consistent with Wheeler’s prior testimony, the video showed KD telling Wheeler that she sucked the defendant’s “peepee” one time and touched it one time, with both KD and Wheeler repeating these statements throughout the interview. Likewise, the video showed Wheeler eliciting from KD, through further questioning and redirection, additional details regarding the alleged fellatio, echoing and expanding upon Wheeler’s testimony to that effect. Lastly, the video showed Wheeler questioning KD about the separate touching incident. KD said that this happened on a different day and with her stepsister, and that the defendant told them both to “quit touching.”
2. THE DEFENDANT’S CASE-IN-CHIEF
As with the prosecution, the defense focused on the credibility of KD’s accounts of the alleged abuse, at*572tempting to undermine their reliability and to impugn Brodie’s motives in connection with them. The defendant first called his mother, who testified that the defendant and KD lived with her for a two-week period and that, during that time, KD slept with her every night and the defendant slept in the office. She further testified that she did not leave her house during that two-week period and that the defendant was never alone with KD there. The defendant’s current wife then testified that the allegations against the defendant came right after they got married and found out they were having a baby. She also testified that Brodie was jealous, was angry with her, and would make constant phone calls to the defendant arguing over KD.
The defendant testified last, and denied the allegations. He testified that on one occasion, when he was living with Brodie and KD was two, he awoke to KD touching his penis when he was sleeping in the nude; he did not know what she touched him with, did not put her in the bed or know how she got there, and would not have slept in the nude if he had known she was going to be there. He “freaked out” and told KD that “it’s a big no, no, you can’t do that.” He then told Brodie, and “there was no big concern about it” because “ [i]t was a two-year-old exploring.” The defendant also testified that, on another occasion, KD and her stepsister came into the bedroom and woke him up by jumping on the bed; he was sleeping in the nude at the time, but was under the covers. The defendant explained that his relationship with Brodie ended “[v]ery badly.” He testified that he initially received custody of KD in the spring of 2008 because the CPS worker investigating the domestic violence charges between him and Brodie concluded that Brodie was the aggressor, and because Brodie had made a statement to the effect that if she could not have KD, no one would. He further testified *573that KD stopped living with him in January 2009 because of issues with Brodie, who would call KD several times a night crying and would tell KD that she did not have to listen to the defendant’s wife. KD was very upset during this time, and so the defendant agreed to let her stay with Brodie to see if that would make things easier on her. The defendant testified that he first learned of the instant allegations of abuse right after he and his wife returned from their honeymoon. He testified that he denied the allegations of abuse when Rothman first interviewed him about them. When Rothman asked him again during their second interview whether KD had performed oral sex on him until he ejaculated, the defendant shook his head no, in disgust; when Rothman then asked whether the defendant could remember that happening, the defendant responded that he could not remember anything like that ever happening.
B. ERRONEOUS ADMISSION OF HEARSAY FROM FORENSIC INTERVIEW
1. ADMISSIBILITY OF HEARSAY UNDER MRE 803A AND MRE 803(24)
We start with the trial court’s admission, over the defendant’s objection, of KD’s out-of-court statements during the forensic interview, which came into evidence through both the testimony of Wheeler and the video recording of that interview. The parties do not dispute that these statements constitute hearsay under MRE 801(c), “offered in evidence to prove the truth of the matter asserted.” The prosecution contends, however, that this hearsay was properly admitted under MRE 803A’s categorical hearsay exception. MRE 803A “codified the common-law ‘tender years exception,’ ” People v Gursky, 486 Mich 596, 607; 786 NW2d 579 (2010), and provides, in relevant part:
*574A statement describing an incident that included a sexual act performed with or on the declarant by the defendant... is admissible to the extent that it corroborates testimony given by the declarant during the same proceeding, provided:
(1) the declarant was under the age of ten when the statement was made;
(2) the statement is shown to have been spontaneous and without indication of manufacture;
(3) either the declarant made the statement immediately after the incident or any delay is excusable as having been caused by fear or other equally effective circumstance; and
(4) the statement is introduced through the testimony of someone other than the declarant.
If the declarant made more than one corroborative statement about the incident, only the first is admissible under this rule.
According to the defendant, KD’s statements to Wheeler during the forensic interview fail to meet many of MRE 803A’s criteria: they were not spontaneously made; they were made over a year after the alleged incidents of abuse, and there has been no showing that this delay was caused by “fear or other equally effective circumstance”; and they do not reflect the first out-of-court statements made by KD corroborating her trial testimony concerning the alleged abuse. Only the last of these challenges was advanced in the defendant’s objection to this evidence at trial, rendering the others unpreserved for our review. See, e.g., People v Aldrich, 246 Mich App 101, 113; 631 NW2d 67 (2001) (“To preserve an evidentiary issue for review, a party opposing the admission of evidence must object at trial and specify the same ground for objection that it asserts on appeal.”). We need not reach these unpreserved issues, however, because we find the defendant’s preserved challenge dispositive.
*575As noted, we will not disturb a trial court’s decision to admit evidence unless that decision “falls outside the range of principled outcomes.” Musser, 494 Mich at 348. Such circumstances are present here. As the defendant argued, and the prosecution conceded before the trial court, KD’s disclosure of the alleged fellatio to Wheeler was not her first corroborative statement regarding that incident; rather, Brodie testified that KD previously disclosed that incident to her, which led to KD’s interview with Wheeler. As a result, MRE 803A does not permit admission of KD’s disclosure of the alleged fellatio during the forensic interview.
The prosecution notes that KD’s disclosure to Wheeler of the separate touching incident was her first corroborative statement to that effect. Even if so,4 it does not render KD’s disclosure of the alleged fellatio to Wheeler any more admissible under MRE 803A, which permits only the first corroborative statement as to each “incident that included a sexual act performed with or on the declarant by the defendant.” Though the statute does not define the term “incident,” it is commonly understood to mean “an occurrence or event,” or “a distinct piece of action, as in a story.” Random House Webster’s College Dictionary (2001). There is no dispute here that the alleged fellatio and touching were distinct occurrences or events, separated by at least a number of months, taking place under different circumstances, and bearing no particular relation to one another be*576yond the parties involved. KD’s disclosure of the fellatio incident to Wheeler does not become admissible under MRE 803A simply because her first disclosure of the touching incident followed shortly after it.5
Accordingly, KD’s disclosure of the alleged fellatio to Wheeler falls outside the plain scope of MRE 803A’s hearsay exception and was improperly admitted under that rule. The prosecution, however, argues on appeal that KD’s out-of-court statements were nonetheless admissible under MRE 803(24)’s residual hearsay exception, citing People v Katt, 468 Mich 272, 290; 662 NW2d 12 (2003), in support. Like the Court of Appeals, we are not persuaded. As this Court has summarized,
To be admitted under MRE 803(24), a hearsay statement must: (1) demonstrate circumstantial guarantees of trustworthiness equivalent to the categorical exceptions, (2) be relevant to a material fact, (3) be the most probative evidence of that fact reasonably available, and (4) serve the interests of justice by its admission. [Id. at 290.]
The requirements of this residual exception “are stringent and will rarely be met, alleviating concerns that [it] will ‘swallow’ the categorical [hearsay] exceptions through overuse.” Id. at 289.
Applying this standard in Katt, this Court held that a child’s disclosure of sexual abuse to a CPS worker, though inadmissible under MRE 803A because it was not the child’s first corroborative statement concerning the abuse, was nonetheless admissible under MRE 803(24). That result is not warranted here. First, KD’s disclosure of the alleged fellatio to Wheeler was not “the most probative evidence of that fact reasonably avail*577able.” Katt, 468 Mich at 290. This is “essentially... a ‘best evidence’ requirement,” which “is a high bar and will effectively limit use of the residual exception to exceptional circumstances.” Id. at 293 (quotation marks and citation omitted). In this case, the “best evidence” of KD’s out-of-court disclosure of the alleged fellatio was the statement made to Brodie prior to the forensic interview with Wheeler. To conclude otherwise would contravene MRE 803A’s express preference for first corroborative statements, and the rationale underlying it. See id. at 296 (“[T]he tender-years rule prefers a child’s first statement over later statements” because, “[a]s time goes on, a child’s perceptions become more and more influenced by the reactions of the adults with whom the child speaks.”). MRE 803(24)’s residual exception cannot be used to “swallow” MRE 803A’s categorical one in this fashion. Id. at 289. The testimony at issue in Katt did not present this same risk; while the child had previously disclosed the abuse to his mother, that first corroborative statement was not available or presented at trial. See id. at 295, 296. Not so here, and nothing in Katt indicates that Wheeler’s testimony regarding KD’s disclosure was properly admitted in addition to Brodie’s.
Similarly, unlike the testimony in Katt, KD’s disclosure to Wheeler does not “demonstrate circumstantial guarantees of trustworthiness equivalent to” those required under MRE 803A, such that it merits admission despite its failure to meet those requirements. “To be admitted, residual hearsay must reach the same quantum of reliability as categorical hearsay; simply, it must do so in different ways.” Id. at 289-290. Thus, if a statement is “deficient in one or more requirements of a categorical exception, those deficiencies must be made up by alternate indicia of trustworthiness,” discerned *578from “the ‘totality of the circumstances’ surrounding [the] statement.” Id. at 289, 291.
Here, Wheeler’s testimony regarding KD’s disclosure of the fellatio incident does not satisfy MRE 803A’s categorical hearsay exception because it was not her first corroborative statement; its spontaneity and delayed nature have also been questioned under that rule. The prosecution notes that the disclosure is nonetheless sufficiently trustworthy under MRE 803(24) because it, like the disclosure in Katt, was made in the course of a properly administered forensic interview. Katt, however, is again distinguishable, and does not support this conclusion. While the disclosure in Katt occurred during a properly administered forensic interview, that interview was intended to address unrelated concerns regarding potential physical abuse by the child’s mother. During the interview, the child spontaneously said that the defendant, his father, did “nasty stuff” to him and then disclosed numerous instances of sexual abuse. No investigation regarding such abuse had begun, and neither the child’s mother nor the interviewer knew that the interview would include this subject. Given the clear spontaneity of the disclosure, the lack of any motive to lie on the part of the mother or child, and the interviewer’s questioning methods, this Court concluded that the disclosure possessed “circumstantial guarantees of trustworthiness equivalent to the categorical exceptions.” Id. at 296.
Similar circumstantial guarantees were lacking here. The specific purpose of Wheeler’s interview of KD was to investigate her prior disclosure of the alleged fellatio — a fact known to both Wheeler and Brodie before the interview — and Brodie’s motives in connection with KD’s disclosure and interview were strongly disputed. Indeed, concern that KD’s statements were *579improperly influenced by Brodie not only animates the defendant’s challenges to their spontaneity and delay under MRE 803A, but also informs their inadmissibility under that rule’s first corroborative statement requirement. See Katt, 468 Mich at 296. While the interviewing methods used by Wheeler may bear on the extent of this concern, we do not conclude, and Katt does not indicate, that they were alone sufficient to cure it. Nor do we see how these methods, or any other circumstances of this case, afforded KD’s disclosure to Wheeler “alternative indicia of trustworthiness” such that it should be deemed any more admissible under MRE 803(24)’s residual exception than it is under MRE 803A’s categorical one.
2. PREJUDICE FROM ERRONEOUSLY ADMITTED HEARSAY
Accordingly, we conclude that the trial court abused its discretion by admitting KD’s out-of-court statements to Wheeler regarding the alleged fellatio. We further conclude that this preserved error more probably than not undermined the reliability of the verdict against the defendant, warranting relief. Musser, 494 Mich at 348. In reaching this conclusion, we consider “ ‘the nature of the error in light of the weight and strength of the untainted evidence.’ “ Id., quoting Krueger, 466 Mich at 54. In particular, as this Court has recognized,
In a trial where the evidence essentially presents a one-on-one credibility contest between the victim and the defendant, hearsay evidence may tip the scales against the defendant, which means that the error is more harmful. This may be even more likely when the hearsay statement was made by a young child, as opposed to an older child or adult. [Gursky, 486 Mich at 620-621 (footnote omitted), citing People v Straight, 430 Mich 418, 427-428; 424 NW2d 257 (1988); People v Smith, 456 Mich 543, 555 n 5; 581 NW2d 654 (1998).]
*580See also People v Anderson, 446 Mich 392, 407 n 37; 521 NW2d 538 (1994) (“While credibility contests are not uncommon in criminal sexual conduct cases, the wrongful admission of corroborating testimony ‘on either side could tip the scales’ and result in harmful error.”), quoting People v Gee, 406 Mich 279, 283; 278 NW2d 304 (1979) (citations omitted).
This case presented the jury with a pure credibility contest; there were no third-party witnesses to either instance of alleged abuse, nor any physical evidence of it.6 As such, the prosecution’s case hinged heavily on KD’s credibility in her accounts of the alleged abuse, particularly the fellatio. With regard to the alleged fellatio, the only accounts properly before the jury were KD’s testimony at trial, and Brodie’s testimony regarding KD’s prior disclosure of it to her.7 The credibility of these accounts, and Brodie’s motives and influence in connection with them, were the focus of the defense and a central issue at trial. As a result of the court’s error, however, the prosecution was not limited to this evidence, and instead the jury was permitted to hear from KD twice more: first, through the hearsay testimony offered by Wheeler, and then again through the video recording of KD’s forensic interview.
The prosecution characterizes this evidence as harmlessly cumulative of KD’s in-court testimony, pointing to our observations in Gursky that “where a hearsay *581statement is not offered and argued as substantive proof of guilt, but rather offered merely to corroborate the child’s testimony, it is more likely that the error will be harmless,” and that “[w]here the declarant himself testifies and is subject to cross-examination, the hearsay testimony is of less importance and less prejudicial.” Gursky, 486 Mich at 620-621. As we also cautioned in Gursky, however, “ ‘the fact that the statement [is] cumulative, standing alone, does not automatically result in a finding of harmless error,’ ” but is only one consideration to be accounted for when evaluating the prejudicial effect of the erroneously admitted hearsay. Id. (citation omitted). Thus, such cumulative hearsay testimony is more likely to be harmless where, unlike here, there is other evidence to corroborate the allegations beyond the declarant’s statements; meanwhile, the likelihood of harm may only increase where, as here, the declarant was a young child and the case was a pure credibility contest. Id.
Based on the evidence presented in this case, we cannot conclude that Wheeler’s testimony and the video recording of the forensic interview were harmlessly cumulative; this hearsay evidence not only corroborated by echo KD’s in-court testimony, but added clarity, detail, and legitimacy to it. KD’s account of the fellatio at trial, while incriminating, left ample room for reasonable doubt; it betrayed uncertainty on fundamental details, was inconsistent in certain respects with Brodie’s corroborative testimony, and was clouded by the strongly disputed motives of Brodie. The evidence of KD’s disclosures to Wheeler, however, did much to alleviate this doubt. Rather than simply Brodie corroborating KD’s testimony, there now too was Wheeler, an expert no less, with no apparent partiality, repeating, clarifying, and more fully articulating KD’s general allegations. The video recording of the forensic inter*582view provided further reinforcement still, as the jury was able to watch KD herself testify again, this time at greater length, with the assistance of Wheeler’s expert questioning, and not subject to cross-examination, of course. This video confirmed Wheeler’s rendition of KD’s statements, repeated them more times over, and elaborated upon them, adding further detail to the graphic scene the prior testimony had sketched.
The resulting prejudice is unsurprising. Wheeler’s testimony and the video recording of KD’s forensic interview left the jury with a much fuller, clearer, and more inculpatory account of the alleged fellatio than that which was properly admitted through KD and corroborated by Brodie. That this elucidation and reinforcement came through Wheeler, presented as a neutral and authoritative source in this pure credibility contest, only heightened the likelihood of its prejudice.
The prosecution contends that any prejudice was immaterial in light of the defendant’s tacit admissions, pointing in particular to his failure to offer an outright denial to Trooper Rothman of the allegations of fellatio, saying instead that he did not remember anything of that sort happening. At trial, the defendant admitted to giving this response, but characterized his choice of words as responsive to Trooper Rothman’s specific question; according to the defendant, when Rothman asked if the fellatio did, in fact, occur, he denied it. While the jury certainly may have factored this testimony into its assessment of the defendant’s credibility, we, like the Court of Appeals, do not find it, or the other untainted evidence offered at trial, sufficiently powerful to restore confidence in the jury’s verdict in light of the trial court’s error. Rather, we conclude that KD’s erroneously admitted statements during the forensic interview more probably than not “tipped the scales” against *583the defendant such that the reliability of the verdict against him was undermined and a new trial is warranted. See, e.g., Gursky, 486 Mich at 621; Straight, 430 Mich at 427-428; Anderson, 446 Mich at 407 n 37.
C. INEFFECTIVE ASSISTANCE OF COUNSEL AT TRIAL
We find this relief likewise warranted by defense counsel’s mishandling of inadmissible testimony offered by Wheeler, Fallone, and Muir vouching for KD’s credibility. As noted, Fallone testified that, based on her investigation, she found that KD’s “allegations had been substantiated” and that, “based on the disclosures made at Care House, there was no indication that [KD] was coached or being untruthful.]” As the Court of Appeals held, this testimony violated the well-established principle that “it is improper for a witness or an expert to comment or provide an opinion on the credibility of another person while testifying at trial.” Musser, 494 Mich at 349. See, e.g., People v Dobek, 274 Mich App 58, 71; 732 NW2d 546 (2007) (“It is generally improper for a witness to comment or provide an opinion on the credibility of another witness, because credibility matters are to be determined by the jury.”). Wheeler likewise violated this principle when she offered her expert conclusions that KD had not been coached by Brodie but rather was being truthful with her. See People v Peterson, 450 Mich 349, 352; 537 NW2d 857 (1995) (affirming that “an expert may not vouch for the veracity of a victim”).8
*584Similarly, Muir testified that, when Brodie confronted the defendant with KD’s allegations, the defendant denied them, leading Brodie to respond that “I know my daughter don’t lie; why is she making these allegations then.” The Court of Appeals found this testimony constituted inadmissible hearsay that improperly vouched for KD’s credibility. There is no dispute that Brodie’s out-of-court statements did not fall under any hearsay exception and, to the extent they were offered for their truth, they were not properly admitted. The prosecution, however, contends that Muir did not offer these statements for their truth, but only to provide context to the defendant’s half of the conversation, which was properly admitted under MRE 801(d)(2)(A).9 Even if so, we do not find Brodie’s commentary on KD’s credibility any more admissible. Muir properly testified to the defendant’s out-of-court denial of the allegations Brodie put to him; Brodie’s statement that “I know my daughter don’t lie” did not provide any meaningful context to this denial and could have easily been omitted “without harming the probative value of *585[the] defendant’s statements.” Musser, 494 Mich at 356. Furthermore, any minimal contextual value this statement may have added was substantially outweighed by the risk that the jury would take the statement for its truth — a risk of particular significance in the context of a case such as this. See id. at 357-358 (explaining that, “especially in child-sexual abuse cases,” “a trial court should be particularly mindful that when a statement is not being offered for the truth of the matter asserted and would otherwise be inadmissible if a witness testified to the same at trial, there is a danger that the jury might have difficulty limiting its consideration of the material to its proper purpose” of providing context to the defendant’s responses) (quotation marks and alteration marks omitted).
Despite the plainly inadmissible nature of the testimony from Fallone and Muir, defense counsel did not object. And while defense counsel initially, and successfully, opposed the prosecution’s attempt to elicit an expert conclusion from Wheeler regarding the veracity of KD’s statements, he thereafter inexplicably permitted that testimony without objection. We agree with the Court of Appeals that, as a result, the defendant was denied the effective assistance of counsel. To be constitutionally effective, counsel’s performance must meet an “objective standard of reasonableness.” Trakhtenberg, 493 Mich at 51. In showing this standard has not been met, “a defendant must overcome the strong presumption that counsel’s performance was born from a sound trial strategy.” Id. at 52, citing Strickland v Washington, 466 US 668, 689; 104 S Ct 2052; 80 L Ed 674 (1984). The strategy, however, in fact must be sound, and counsel’s decisions as to it objectively reasonable; “a court cannot insulate the review of counsel’s performance by calling it trial strategy.” Id.
*586We see no sound strategy in counsel’s failure to object to the vouching testimony offered by Wheeler, Fallone, and Muir. As defense counsel affirmed at the Ginther hearing, his trial strategy was to demonstrate that KD was not believable, that her testimony had been tainted by Brodie, and that she had told different stories to different people throughout the investigative process. In fact, he also testified that, consistent with this strategy, he would have objected to any opinions offered that KD was being truthful. Wheeler’s and Fallone’s testimony that KD was telling the truth, and Muir’s recounting of Brodie’s statements to that same effect, directly contravened this strategy. Defense counsel offered, and we see, no strategic reason to permit this inadmissible testimony to pass without objection here.10
We further conclude that, but for these deficiencies in counsel’s performance, “there is a reasonable probability that the outcome of [the defendant’s trial] would have been different.” Trakhtenberg, 493 Mich at 51. As already discussed, the prosecution’s case hinged wholly on the credibility of KD’s allegations, making defense counsel’s success in undermining that credibility all the more critical. Rather than pursuing this strategy vigilantly, defense counsel permitted Wheeler, Fallone, and Muir — three figures of apparent authority and impartiality, with direct involvement in and knowledge of the investigation leading to the defendant’s prosecution — to present testimony improperly reaching the key factual issue before the jury: whether ED was telling the truth. Wheeler’s and Fallone’s commentary was especially prejudicial in this regard — the former *587offering the jury an expert opinion regarding KD’s credibility in the instant case, and the latter offering the jury her, and CPS’s, professional assessment of the veracity and substantiation of KD’s complaints. We cannot overlook the influence such testimony may have in a case such as this. See Musser, 494 Mich at 357-358 (noting that, “given ‘the reliability problems created by children’s suggestibility,’ ” this Court “has condemned opinions related to the truthfulness of alleged child-sexual-abuse complainants” because the jury in such credibility contests “is often ‘looking to “hang its hat” on the testimony of witnesses it views as impartial’ ”), quoting Peterson, 450 Mich at 371, 376.11 Given the *588centrality of KD’s credibility to the prosecution’s case, the lack of evidence beyond her allegations, and the nature of the testimony offered by Wheeler, Fallone, and Muir, we believe it reasonably probable that, but for this testimony, the outcome of the defendant’s trial may have been different. See Musser, 494 Mich at 363-364.12
*589D. CONCLUSION
We thus conclude that the defendant is entitled to a new trial as a result of both the trial court’s erroneous admission of KD’s statements regarding the alleged fellatio during her forensic interview, and defense counsel’s ineffective assistance with respect to the testimony of Wheeler, Fallone, and Muir. This case put before the jury serious and disturbing allegations, heavily contested facts and motives, and a singular, difficult choice: whether to believe KD or the defendant. The trial court’s and defense counsel’s errors each bore directly and significantly upon this choice. For the reasons discussed, we find the prejudicial effect of each of these errors too strong, and the untainted evidence too weak, to conclude that the jury’s verdict against the defendant remains sufficiently reliable to stand. We therefore affirm the Court of Appeals’ conclusion that the defendant is entitled to a new trial.
Because we do not find them necessary to this award of relief, we do not reach a number of the defendant’s unpreserved evidentiary challenges: namely, whether KD’s disclosures of the alleged touching and fellatio incidents to Wheeler were inadmissible under MRE 803A because they were not spontaneously made, as well as whether those disclosures, and KD’s disclosure of the alleged fellatio to Brodie, were inadmissible under that rule because there was no demonstration of “fear or equally effective circumstance” excusing their substantial delay. The parties remain free to litigate these issues on retrial. We take this opportunity to note, however, that we agree with the observations in Judge RONAYNE Krause’s concurring opinion in the Court of Appeals that, when evaluating whether a delay in *590disclosure is excusable under MRE 803A, courts should bear in mind that “MRE 803A(3) requires any circumstance that would be similar in its effect on a victim as fear in inducing a delay in reporting, not a circumstance that is necessarily similar in nature to fear,” and that “[n]othing in the rule even requires that any ‘other equally effective circumstance’ must have been affirmatively created by the defendant.” Douglas, 296 Mich App at 211 (RONAYNE KEAUSE, J., concurring).13 We need not set forth a list of circumstances that are similar to fear in their effect on a child, as the determination whether such circumstances exist should be done by the trial court on a case-by-case basis. We likewise express no opinion as to whether such circumstances are present in this case — indeed, we agree with the sentiment expressed by Judge RONAYNE KRAUSE and shared by the Court of Appeals majority that the present record is “disappointing” in that regard — but leave the development and determination of that issue to the trial court in the first instance, if and when the issue is put before it.
m THE DEFENDANT’S ENTITLEMENT TO REINSTATEMENT OF PLEA OFFER
While we agree with the. Court of Appeals that the defendant is entitled to a new trial, we disagree that he is entitled to relief on the basis of his counsel’s deficient performance at the pretrial stage. Although during the *591plea-bargaining process counsel indisputably misadvised the defendant of the consequences he faced if convicted at trial, the trial court did not reversibly err in determining that the defendant has not shown prejudice as a result of counsel’s deficient performance.
Before trial, the defendant was presented with two plea offers: the first, made before the preliminary examination, was for the defendant to plead guilty to attempted CSC, carrying a five-year maximum penalty; the second, made just before trial, was for the defendant to plead guilty to CSC-iy carrying a two-year maximum penalty. As to the first offer, counsel advised the defendant that the plea would likely entail jail rather than prison time; as to the second, that the defendant would serve ten months in county jail and would have to register as a sex offender. The defendant rejected both offers. There is no disagreement that counsel never informed the defendant that he faced a 25-year mandatory minimum prison sentence if convicted of CSC-I at trial. See MCL 750.520b(2)(b). Instead, counsel mistakenly advised the defendant that a conviction at trial would result in a potential maximum sentence of 20 years in prison, and that he would likely have to serve approximately five to eight years before being eligible for parole. Counsel also informed the defendant that a conviction for any CSC offense would require that he register as a sex offender.
According to the defendant, counsel’s failure to properly advise him of the 25-year mandatory minimum sentence, as well as of certain consequences of sex-offender registration, denied him the effective assistance of counsel; as a result, the defendant contends, he is entitled to reinstatement of the prosecution’s second plea offer. As at trial, a defendant is entitled to the effective assistance of counsel in the plea-bargaining *592process. Lafler v Cooper, 566 US_, _; 132 S Ct 1376, 1384; 182 L Ed 2d 398 (2012). A defendant seeking relief for ineffective assistance in this context must meet Strickland’s familiar two-pronged standard by showing (1) “that counsel’s representation fell below an objective standard of reasonableness,” and (2) “that there is a reasonable probability that, but for counsel’s unprofessional errors, the result of the proceeding would have been different.” Id. at_; 132 S Ct at 1384. In demonstrating prejudice, the “defendant must show the outcome of the plea process would have been different with competent advice.” Id. at_; 132 S Ct at 1384. Where, as here, the alleged prejudice resulting from counsel’s ineffectiveness is that the defendant rejected a plea offer and stood trial,
a defendant must show that but for the ineffective advice of counsel there is a reasonable probability that the plea offer would have been presented to the court {i.e., that the defendant would have accepted the plea and the prosecution would not have withdrawn it in light of intervening circumstances), that the court would have accepted its terms, and that the conviction or sentence, or both, under the offer’s terms would have been less severe than under the judgment and sentence that in fact were imposed. [Id. at_; 132 S Ct at 1385.]
The defendant has the burden of establishing the factual predicate of his ineffective assistance claim. People v Hoag, 460 Mich 1, 6; 594 NW2d 57 (1999). And as already noted, a trial court’s factual findings in that regard are reviewed for clear error and cannot be disturbed unless “the reviewing court is left with a definite and firm conviction that the trial court made a mistake.” Armstrong, 490 Mich at 289. See MCR 2.613(C).
*593Here, after hearing testimony from the defendant and defense counsel at the Ginther hearing, the trial court rejected the defendant’s claim of ineffective assistance. The court found that the defendant, at the time he rejected the prosecution’s second plea offer, believed that a conviction at trial would result in a 20-year maximum prison sentence, supervised to no contact with his children for 20 years, and registration as a sex offender. The court reasoned that, although the defendant thought he faced a 20-year maximum sentence rather than a 25-year mandatory minimum one if convicted, this misinformation made “no difference” in light of the defendant’s proclamations of “his innocence in the face of plea bargains that were offered.” The Court of Appeals reversed this determination, finding counsel’s performance deficient and rejecting the trial court’s conclusion that the defendant did not suffer prejudice as a result, explaining that (1) “there is a significant difference between the possibility of a 20-year term with the likelihood of serving a much shorter sentence and the certainty of serving a 25-year minimum term”; (2) defense counsel testified at the Ginther hearing that he would have “absolutely pressed” the defendant to accept the plea had counsel known of the 25-year mandatory minimum at the time; (3) the defendant likewise testified that he would have accepted that plea with the correct sentencing information, even if it meant limited to no contact with his children; and (4) the defendant testified that that his decision to reject the plea offer was also affected hy counsel’s mistaken advice that he would not be permitted to reside with his children for as long as he was required to register as a sex offender.
We agree with the Court of Appeals that counsel’s mistaken advice regarding the sentence the defendant faced at trial fell below an objective standard of reason*594ableness.14 See, e.g., Padilla v Kentucky, 559 US 356, 370; 130 S Ct 1473; 176 L Ed 2d 284 (2010) (noting “the critical obligation of counsel to advise the client of the advantages and disadvantages of a plea agreement”) (quotation marks omitted); People v Corteway, 212 Mich *595App 442, 446; 538 NW2d 60 (1995) (explaining that counsel must provide advice during plea negotiations that is sufficient to allow the defendant “to make an informed and voluntary choice between trial and a guilty plea”). We likewise agree that the difference between a required 25-year mandatory minimum sentence and a possible 20-year maximum one might potentially affect an individual’s decision whether to accept a plea or go to trial. The trial court, however, did not conclude otherwise; rather, it found that this difference would not have affected this particular defendant’s decision to reject the pleas in this case in light of his protestations of innocence. We do not see reversible error in this determination.
In concluding otherwise, the Court of Appeals made no mention of the role that the defendant’s belief in his innocence may have played in his decision to go to trial, despite its prominent place in the trial court’s reasoning, and instead focused on certain testimony offered by defense counsel and the defendant that knowledge of the 25-year mandatory minimum would have affected their treatment of the prosecution’s plea offer. Review of that testimony in full, however, paints a different picture. First, contrary to the Court of Appeals’ characterization, defense counsel did not testify that he would have “absolutely pressed” the defendant to accept the prosecution’s plea offer had he known of the 25-year mandatory minimum at the time. Rather, counsel stated that, “ [i]f there was a do-over on this, I would have absolutely pressed [the defendant] and insisted he take the deal. . . because we lost at trial, and the consequences are he’s now looking at 25 years in prison.” When asked what he would have done differently had he only known about the mandatory minimum, however, and not the ultimate outcome of the trial, defense counsel was much more equivocal in his *596response, saying simply that he “would have made sure [the defendant] understood how long 25 years was.” Counsel further testified that his and the defendant’s position had always been that the defendant would plead to nothing that would result in placing the defendant on the sex-offender registry, in part because the defendant was concerned about losing contact with his children, but also because he found the type of behavior to which he would be pleading “disgusting and offensive and [he] would never engage in” it.15 Correspondingly, defense counsel testified that the defendant has always maintained his innocence, a claim that defense counsel believed. This is consistent with defense counsel’s earlier representation to the trial court at the defendant’s sentencing hearing, in which he indicated that the defendant has “made it perfectly clear,” from arraignment and “consistently since,” that “he did not commit this crime,” and that the defendant “has made it clear that he turned down numerous plea bargains because he was basing his decision . . . upon his innocence.”
Meanwhile, the defendant, as the Court of Appeals noted, testified that he would have accepted a plea had he known of the 25-year mandatory minimum, and also suggested that he would have been more inclined to accept a plea had he not mistakenly believed that sex-offender registration would prohibit him from living with his children for its duration. As noted, it is questionable that the defendant’s misconceptions regarding the consequences of sex-offender registration were caused by any deficient performance on counsel’s *597part. In any event, the full body of the defendant’s testimony undermines the credibility of his assertions that either these misconceptions or the misinformation regarding the sentence he faced at trial meaningfully influenced his decision to reject the prosecution’s plea offer. For instance, the defendant testified that the only way he would have taken a plea was if he knew of the 25-year mandatory minimum, and that he still would have taken the plea even if it meant limited to no contact with his children for a period of time. He also testified, however, that he would not have accepted any plea that required sex-offender registration because he was innocent and because it would affect his relationship with his children. The defendant further testified that he probably would not have accepted a plea that required any jail time and that, in deciding to reject the prosecution’s plea offer, the minimum sentence he faced at trial did not matter because he was innocent, he did not commit the crime, and he did not think he would lose. This testimony is confusing at best, and casts significant doubt upon what circumstances, if any, would have led the defendant to accept a plea. It certainly betrays no clear error in the trial court’s discernment of the common thread running throughout both the defendant’s and his counsel’s testimony: that the defendant rejected the prosecution’s plea offers because he was innocent of the charges, was not a sex offender, and was not interested in pleading guilty to repugnant acts that he did not commit.16
As a result, we are not “left with a definite and firm conviction that the trial court made a mistake” in *598finding that the defendant has failed to show prejudice stemming from his counsel’s deficient performance, Armstrong, 490 Mich at 289; rather, the record amply supports the conclusion that, even had the defendant been properly advised of the consequences he faced if convicted at trial, it was not reasonably probable that he would have accepted the prosecution’s plea offer. See Lafler, 566 US at_; 132 S Ct at 1384-1385.17 There is *599no indication that the trial court failed to duly consider the record in making its determination, including the terms of the plea available to the defendant, the consequences the defendant faced in rejecting that plea, the defendant’s understanding of the plea and those consequences, and the defendant’s motivations for assuming the risks of trial. Because we see no clear error in the trial court’s factual findings, nor any legal error in its analysis, we find no basis to reverse the trial court’s conclusion that relief is not warranted for the defendant’s claim of ineffective assistance at the pretrial stage. We reverse the Court of Appeals on this point and hold that the defendant is not entitled to reinstatement of the prosecution’s plea offer.18
V CONCLUSION
For the foregoing reasons, we conclude that the defendant is entitled to a new trial, but is not entitled to reinstatement of the prosecution’s plea offer. Accordingly, we affirm the judgment of the Court of Appeals in part, reverse in part, and remand for proceedings consistent with this opinion. In addition, we deny as moot the defendant’s motion to expand the record.
YOUNG, C.J., and Kelly and Zahra, JJ., concurred with McCormack, J.People v Ginther, 390 Mich 436, 212 NW2d 922 (1973).
Judge Amy Ronayne Krause issued a concurring opinion, agreeing with all but one of the majority’s conclusions. Based on the record before the trial court, she concluded that the admission of Brodie’s testimony concerning KD’s initial disclosure did not provide a basis for relief.
Defense counsel asked this question during his brief cross-examination of KD. At the prehminary examination, KD initially testified that the “milk” that came out of the defendant’s “peepee” tasted like cherry, but then said it tasted like “regular milk.” Defense counsel did not otherwise refer to or use KD’s preliminary examination testimony during his examination of her at trial.
The defendant contends that this was not, in fact, KD’s first disclosure of the alleged touching, and seeks to expand the record to support this claim. The defendant, however, did not raise this challenge below, and while KD’s trial testimony suggests that she may have discussed this incident with her therapist, the present record contains no further information to that effect. For the purposes of resolving the instant appeal, we need not reach this dispute, but the parties remain free to litigate it on retrial.
Likewise, KD’s disclosure of the touching incident to Wheeler does not hecome any less admissible under MRE 803A simply because her disclosure of the alleged fellatio incident fails that rule’s first-corroborative-statement requirement.
Of course, such corroborative evidence is not necessary for the defendant to be convicted of the charged offenses, see MCL 750.520h, but its absence is properly considered when evaluating the prejudicial effect of the court’s erroneous admission of KD’s out-of-court statements to Wheeler. See Musser, 494 Mich at 363.
We assume for the purposes of this analysis, without deciding, that Brodie’s account of KD’s disclosure was properly admitted. This in no way forecloses the defendant’s ability to challenge its admissibility on retrial.
The prosecution claims this testimony was no different than testimony that KD’s behavior was consistent with that of a victim of sexual abuse, and thus was properly admitted under Peterson. We disagree. In Peterson, this Court recognized that an expert may offer testimony that a particular child’s specific behavior is consistent with that of a sexually abused child if the defendant either “raises the issue of the particular child victim’s post-incident behavior” or “attacks the child’s credibility” *584by “highlight[ing] behaviors exhibited by the victim that are also behaviors within [the child sexual abuse accommodation syndrome] and allud[ing] that the victim is incredible because of these behaviors.” Peterson, 450 Mich at 373-374 & n 13. Correspondingly, the scope of such expert testimony is limited to the specific behavior at issue. Id. at 374 n 13. Setting aside that Fallone did not testify as an expert, neither her nor Wheeler’s testimony fits these criteria. The defendant had not put at issue or attacked KD’s credibility on the basis of any particular behavior contemplated in Peterson, nor was Wheeler or Fallone explaining how any such behavior was consistent with that of an abused child. Rather, they directly and conclusively opined that KD’s allegations in the instant case were true and trustworthy. Such testimony does not fall not within the carefully circumscribed circumstances identified in Peterson, but instead remains subject to the general prohibition on testimony “vouching] for the veracity of a victim,” which Peterson also affirmed. Id. at 352.
MRE 801(d)(2)(A) provides, “A statement is not hearsay if. .. [t]he statement is offered against a party and is . . . the party’s own statement . . ..”
The trial court, for its part, did not address this evidence when rejecting the defendant’s claim of ineffective assistance of counsel at trial; rather, the only finding it made as to defense counsel’s trial performance was a brief reference to the decision whether to call KD’s stepsister as a witness for the defense.
According to the prosecution, no prejudice inured from Fallone’s testimony because it stated nothing beyond what could be obviously inferred from her presence as a witness for the prosecution, citing Dobek, 274 Mich App at 71, in support. We disagree with this proposition, and do not read Dobek to support it. In that case, a police officer testified that he had no concern that the child-complainant was lying in her statements to him regarding the alleged sexual abuse at issue; this testimony was not objected to and occurred on redirect examination, after defense counsel had asked on cross-examination whether the officer’s observation that the complainant appeared uncomfortable was consistent with how individuals who are lying may appear. The defendant claimed prosecutorial misconduct in eliciting this testimony. The court of appeals rejected that claim, as it could not “conclude that the prosecutor proceeded with the questioning and elicited the testimony in bad faith, especially considering that defendant opened the door on the matter.” Id. The court additionally noted that, “[ajssuming plain error [in the testimony’s admission], defendant has not established prejudice, actual innocence, or damage to the integrity of the judicial proceedings” as to this unpreserved error because, “[g]iven that [the officer] was called as a witness by the prosecutor and that a criminal prosecution against defendant was pursued, the jurors surely understood that [the officer] believed that the victim was telling the truth even without the disputed testimony.” Id.
Dobek thus held that certain erroneously admitted vouching testimony did not warrant relief because it was elicited in direct response to defense counsel’s questioning on the topic and was reviewed for prejudice *588under a significantly more deferential standard than is applicable here. Dobek does not suggest that Fallone’s mere presence on the stand as a witness for the prosecution cures any prejudice caused by her testimony vouching for KD, nor does it cast doubt upon our conclusion that, but for counsel’s ineffectiveness as to the testimony of Wheeler, Fallone, and Muir, there was a reasonable probability that the outcome of the defendant’s trial would have been different.
In addition to the mishandling of Wheeler’s, Fallone’s, and Muir’s testimony, the Court of Appeals also found defense counsel ineffective for failing to impeach KD with certain inconsistencies between her trial testimony and her preliminary examination testimony, noting that there was “no logical reason” for not doing so. We disagree. As defense counsel explained at the Ginther hearing, his strategy with KD at trial, as a very young and sympathetic witness, “was to get her on and off the stand as quick as possible.” Defense counsel made a similar point to the jury during his closing argument, explaining that he did not press KD on the stand regarding certain details because “[tjhere’s just things that a child doesn’t need to go through, and there’s just no good way to do things like that.” Furthermore, although, as the dissent observes, certain portions of KD’s preliminary examination testimony were potentially damaging to her credibility, other portions were potentially supportive of it, corroborating her trial testimony. Thus, while we agree with the Court of Appeals that KD’s preliminary examination testimony contained material with which defense counsel could have attempted to impeach her at trial, we find it objectively reasonable for him to have concluded, given the circumstances, that the risks of this attempt outweighed its potential benefits.
The Court of Appeals also concluded that defense counsel performed deficiently by failing to object to KD’s disclosures to Brodie and Wheeler on the basis of their delayed nature. Just as we need not reach the merits of that objection here, our disposition of this appeal does not require us to determine whether counsel was ineffective for failing to pursue it. We note, however, that the timing of KD’s disclosures supported the defendant’s theory that Brodie fabricated them out of spite; defense counsel thus may have chosen not to object to KD’s disclosures on the basis of delay so as not to encourage the development of a record at trial that might provide alternate explanations for that delay. We are thus not *589convinced that defense counsel’s failure to object on this basis was constitutionally ineffective, given how it dovetailed with his trial strategy.
Indeed, prior to the enactment of the rules of evidence, this Court recognized that circumstances similar to fear in their effect on a child were sufficient to excuse a delayed disclosure under the common-law “tender years exception,” which, as previously noted, MRE 803A codified. See People v Baker, 251 Mich 322, 326; 232 NW 381 (1930) (finding delay in the child’s disclosure of abuse by her father excused in that case because the abuse coupled with the father’s “admonition to her not to tell [were] as effective to promote delay as threats by a stranger would have been”).
The Court of Appeals did not expressly determine whether counsel performed deficiently in advising the defendant of the consequences of sex-offender registration; it did, however, note that “[although defense counsel advised defendant in this case that he would be required to register as a sex offender, counsel erroneously informed defendant that his registration would preclude him from living with his children for the duration of his registry, or 20 years.” Douglas, 296 Mich App at 208 n 6. In light of our determination that the defendant has not carried his burden of showing prejudice as a result of counsel’s claimed errors at the pretrial stage, we need not reach this question here. We note, however, that the record leaves considerable doubt as to whether counsel provided ineffective assistance regarding this aspect of his advice. According to the defendant, counsel incorrectly advised him that such registration would preclude him from living with or seeing his children for 20 years, a consequence which he believed would attach regardless of whether he accepted a plea or went to trial. According to defense counsel, however, he advised the defendant that sex-offender registration was mandatory for 25 years and he “had several conversations with [the defendant] . .. about the terrible things that happenQ to somebody that’s on the sex offenders’ list as it relates to their relationship with their children.” Namely, counsel advised the defendant that “his relationship with his children would be severely jeopardized with a CSC conviction,” that “his contact with his children could be severely restricted, particularly if. .. he was on probation or parole,” and that if “he was convicted in Lenawee County of a CSC charge involving one of his daughters, it was going to be very difficult to achieve any type of visitation with any of his other children” as there was a “real likelihood” that CPS would open an investigation and restrict the defendant’s ability to live or have unsupervised contact with any of his three children (including KD, but also the daughter he had with his wife and his wife’s daughter from a prior relationship). While defense counsel indicated a connection between these consequences and the defendant’s registration as a sex offender, it is not apparent that he advised the defendant that they arose directly from or would necessarily last the duration of that registration. Nor is it apparent that he otherwise misadvised the defendant as to the nature of these consequences, their relationship to sex-offender registration, or how they might inform the defendant’s decision whether to accept a plea or go to trial.
Defense counsel also testified that he told the defendant that pleading guilty would require an admission of the acts to which he pled and also completion of sex offender therapy, which would likewise require such an admission.
The prosecution urges that the defendant cannot show prejudice as a matter of law in light of his maintenance of innocence, because Michigan does not authorize the acceptance of guilty pleas under such circumstances. Our analysis here neither reaches nor endorses this position; rather, we simply conclude that, under the facts of this case, the *598trial court did not clearly err in evaluating how this particular defendant’s belief in his innocence affected his decision to reject the plea offers put before him.
The dissent raises two primary objections to this conclusion, neither of which we find convincing. First, the dissent avers that we have “mischaracteriz[ed the] defendant’s posttrial testimony” in our analysis and that the defendant, as reflected by select portions of that testimony, “consistently maintained that he would have responded differently to the prosecution’s offers if he had known about the mandatory minimum sentence he was facing.” Simply put, the dissent finds a coherence in the defendant’s assertion of prejudice that our review of his testimony in full, along with that of his counsel, does not support. We fail to see any mischaracterization in our summary of that testimony, or any consistency in it that we have overlooked. The dissent may find the defendant’s testimony more credible than the trial court did, but that of course is not the relevant inquiry. We review the trial court’s factual findings for clear error, and in so doing must give due “regard ... to the special opportunity of the trial court to judge the credibility of the witnesses who appeared before it.” MCR 2.613(C). Neither the record nor the dissent’s impression of it reveals any basis for disregarding this “special opportunity” and finding clear error in the trial court’s evaluation of the defendant’s testimony.
Similarly, the dissent believes that we have “give[n] too little weight to the magnitude of defense counsel’s error” in our analysis, because “[e]ven the most stubborn defendant would at least consider pleading guilty upon learning that he was about to stand trial on a charge for which the statutory minimum sentence was 25 years in prison.” We share the dissent’s appreciation of the magnitude of defense counsel’s error in this case, and likewise recognize the influence such an error might have on an individual’s decision whether to accept a plea of the sort offered here. At issue, however, is the effect of counsel’s error on this particular defendant, not some hypothetical one. This question, as the dissent observes, is inherently counterfactual, but nonetheless one on which the *599trial court can and did ably develop a complete record, and we see no reversible error in the court’s assessment of it. We cannot agree with the dissent that this assessment should have instead been dictated by the trial court’s — or our own — abstract belief of what “[e]ven the most stubborn defendant” might have done, “no matter how [the] defendant actually behaved” in this case.
This, of course, does not foreclose the prosecution from choosing to reoffer this or another plea to the defendant on remand.