Arthur, J.
*175Following a three-day jury trial in the Circuit Court for Anne Arundel County, appellant Joseph Walter was convicted of sexual abuse of a minor, but acquitted of second-degree rape and of committing a third-degree sexual offense. The court sentenced Walter to 20 years' imprisonment and required him to register as a sex offender. He noted this timely appeal.
Because we conclude that the court erred in allowing the jury to view the portions of a recorded interview in which a police detective repeatedly expressed her disbelief of Walter's claims of innocence and her opinions concerning his guilt, we reverse the conviction and remand for a new trial. We address several other matters to provide guidance on remand.
FACTUAL AND PROCEDURAL HISTORY
M.'s Testimony
In the summer of 2012, M., then 12 years old,1 lived with her two half-brothers, her mother ("Mother"), and her stepfather ("Stepfather"). While M., her younger brothers, and Mother visited family members for a month, Walter, who is Stepfather's *54brother, began living in M.'s family home.
M. and the other family members returned to Maryland shortly before school started. M. testified that she was "very *176close" to Walter at that time and that she would spend time together with him and his son.
According to M., one night while Walter was staying at the house, she had a nightmare. She asked Walter to come into her room. They lay down in her bed. He was "probably" wearing boxers; she was wearing sweat pants. She told him that she was going to bed, and he left. He did not touch her inappropriately.
At some point thereafter, M. testified, Walter came into her room again, at night, without an invitation, when the door was closed. He lay down behind her and touched her breast with his hand, outside of her clothing. He said nothing and got up and left.
The next night, she testified, Walter came into her room again, without an invitation, when the door was closed. He lay behind her and touched her breast and her buttocks, under her clothing. He got up and left again, saying nothing.
The following night, M. testified, Walter came into her room yet again, without an invitation, when the door was closed. He touched her breast and buttocks over her clothes while she was lying on her side in bed. At some point, she got to her feet, and he put his penis inside of her while she was standing. He ejaculated into a towel that came from the bathroom.
The next night, M. testified, Walter tried to come into her room again. In describing what happened next, M. testified: "I push him away and tell my mother." It is unclear from M.'s testimony what she told Mother, but a subsequent question implies that she said that Walter was trying to get into her room. The following day, Mother replaced the doorknob on M.'s door with a doorknob that had a lock. The next day, M. testified, Walter moved out of the house.
After Walter moved out, M. continued to see him, go to his house, and hang out with his son. She testified that she said nothing more about what Walter had done because she did not want to tear apart the family.
*177M. first told her parents about the abuse in early 2016.2 She had been out past curfew and was brought home by a police officer, who insisted on speaking with her father. She was very upset, was crying, and was afraid that she was going to get into trouble. At that moment, she told Stepfather that Walter "had sex with [her]." In her testimony, M. admitted that she had lied to her parents to get out of trouble.
M. testified that she and Stepfather waited a day before telling Mother about the abuse. At some point thereafter, they called the police. M. was punished for staying out past curfew.
On cross-examination, the defense elicited testimony suggesting that M. might not be an entirely reliable narrator. M. had initially told the police detectives that Walter had had sexual intercourse with her on two occasions, but at trial she admitted that that was "not true." She had falsely told the detectives that, on one occasion when Walter tried to enter her room, she had pushed back against the door, screamed, and woke up her mother, who had pushed Walter away. She had told Stepfather that Walter choked her, but at *55trial she denied having done so. At first, she had told the detectives that she was "terrified" of Walter and "didn't really talk to him"; in a later interview, after asking whether she could call Walter, she told the detectives that she had already called him and that he had told her that she was lying; but at trial she admitted that she had invented her account of the call and that no such call ever took place.3 *178The cross-examination of M. also suggested that she was having a great deal of trouble at school and at home when she first claimed that Walter had sexually abused her. She had been suspended from the eighth grade, apparently for bringing alcohol and a bong to school, and her grades were not good. She had been in trouble with the police, and she was having a hard time getting along with both of her parents. Her mother was at wit's end. She had been home-schooled for a while, and there was some discussion about sending her to school out-of-state. At the recommendation of Walter's wife, M. received counseling. But although the counseling was confidential, M. did not disclose the abuse to her counselor.
The cross-examination established that M. continued to have contact with Walter after the abuse. She would see him at family gatherings. She called him, sent him text-messages, and talked to him on Facetime, and she went to his house and spent the night there on multiple occasions. She told Walter's wife, with whom she had become close, that she wanted to move out of her parents' house and live with her biological father.
Finally, a few days or weeks before she accused Walter of abusing her, Stepfather gave Walter a used pickup truck that was supposed to be for her.
Mother's Testimony
Mother testified that before 2012 M. was doing well in school, played sports, and had a good group of friends.
During the late summer or early fall of 2012, Mother testified, M. came to her and her husband and said that Walter "was going in and trying to wake her up in the middle of the night." Because M.'s doorknob did not have a lock, they took the doorknob from their bedroom door (which had a lock) and put it on hers. A day or two later, M. told them that Walter "was trying to get in after the locks were changed."
*179The next day, they asked Walter to leave. Walter had told the police detectives that he had been sleepwalking while he lived at his brother's house, but Mother said that she never saw him sleepwalk.
After Walter moved out, Mother testified, M. started to go downhill. Nonetheless, M. continued to have good interactions with Walter during that time, and she would sometimes go to his house.
Mother agreed that M. sometimes lies, but said that she does not lie to get others in trouble. Mother said that M. is punished when she lies.
Mother testified that, in January or February of 2016, M. disclosed the abuse.
*56M. came to her and asked if she "knew." Mother responded, "I'm not quite sure what you're talking about." M. explained that Walter "had slept with her." According to Mother, M. was "very distraught."
After M. disclosed the abuse, Mother conferred with her husband about what to do. After three or four days, she said, they called the police.
On cross-examination, Mother testified that it was actually about four weeks before they spoke to the police detectives and that the delay occurred, in part, because M. was known to lie a lot. In addition, Mother confirmed M.'s testimony about the turmoil in her private life at that time: she had been suspended from school, she had run away, she had had multiple encounters with the police, and her parents had considered sending her out-of-state to a boarding school. Finally, Mother testified that, contrary to M.'s statement to the detectives, she had never heard M. screaming in the night and had never confronted Walter and pushed him away from M.'s door.
Stepfather's Testimony
Stepfather testified that, in the summer or fall of 2012, Walter had a "normal" relationship with M., and M. was a "normal" child. Stepfather said that he never saw any evidence *180of Walter sleepwalking before or during the summer or fall of 2012.
As a result of a conversation with his wife, Stepfather changed the doorknob on M.'s bedroom door about a month after his wife and children returned from their vacation. On the morning after he changed the doorknob, Stepfather had another conversation with his wife. As a result of that conversation, he and his father told Walter that he had scared M. and that he had to leave. Over an objection, Stepfather testified that Walter looked as though he was "trapped" during that conversation. Walter moved out, but continued to come to Stepfather's house to visit.
After Walter moved out, Stepfather testified, M.'s behavior began to change, but he did not elaborate on how. Like his wife, Stepfather agreed that M. sometimes lies, but does not lie to get others in trouble.
In February of 2016, M. had run away. The police brought her home. Stepfather testified that he confronted her, asking "Where is this behavior coming from?," and "[A]re you okay?" She began crying and disclosed the abuse.
Stepfather did not tell his wife what M. had said. Instead, a day or two later, he said, M. told her mother of the abuse.
According to Stepfather, he and his wife called the police about a month later. In the interim, they had taken M. to a counselor. The counselor directed them to call the police.
At the request of the police, Stepfather called Walter. Unbeknownst to Walter, the police were listening in on the call, which was played for the jury. In the call, Stepfather told Walter that, according to M., he had "fingered her" and "choked her out and had sex with her."4 Walter categorically denied the allegations. He claimed that he sleepwalked into her room one night, lay down, woke up, and left the room. In *181addition, he remembered an incident when M. came to the couch where he was sleeping and lay down with him because *57she was having a nightmare.5
On cross-examination, Stepfather could not recall telling the police that he had changed the locks because M. had said that Walter was sleepwalking. Nor could he recall telling the police that Walter had sleepwalked when he was a child.
Testimony of Erin Lemon
Over a defense objection, the State called Erin Lemon, whom the court accepted as an expert "in the fields of sexual child abuse generally, delayed disclosure [of sexual abuse], and child development."
Ms. Lemon, a social work supervisor with the Anne Arundel County Department of Social Services, testified about factors that may cause a child to delay a report of sexual abuse. They include the child's age and development; the child's relationship with the alleged perpetrator; the "grooming process," in which a perpetrator subtly persuades or manipulates the child not to disclose the abuse; and the child's own understanding of what sexual abuse is. On the basis of her experiences, Ms. Lemon opined that "it is very common for children who have been sexually abused to delay reporting that abuse, especially when that abuse ... was caused by a family member or someone who's in a trusted position, known to the family." She did not opine that M. was the victim of sexual abuse or about the reasons for the delayed disclosure in M.'s case.
On cross-examination, Ms. Lemon agreed that she knew nothing about this specific case: she had not interviewed M., her parents, the detectives who investigated the case, or anyone else. She also agreed that a delayed disclosure may be a fabrication; that a prompt disclosure, too, may be a fabrication; and that a child may be fabricating an allegation of *182sexual abuse when he or she claims to have delayed reporting it for fear of breaking up the family.
Ms. Lemon had no data that organized or tabulated her experiences; she could identify no methodology for collecting any information to support the conclusions that she drew from her experiences; and her conclusions had not been published or subjected to peer review.
The cross-examination concluded with Ms. Lemon's concession that the timing of a child's report of abuse is not "diagnostic" of whether it actually happened. There is no direct correlation, she appeared to say, between the timing of the disclosure and the veracity of the complaint.
Testimony of Detective Christine Mays
As its final witness, the State called Detective Christine Mays, who had investigated M.'s allegation of abuse. Detective Mays had listened to the call between Stepfather and Walter, and she had interviewed Walter when he voluntarily came to speak with her at her request.
Over a defense objection, the State played a largely unredacted video-recording of Detective Mays's interview with Walter.6
In the interview, which spans 61 pages, Walter repeatedly denied all of M.'s allegations. He remembered that his brother had changed the lock on M.'s door, but said that he did not know why. He said that he had been sleepwalking because he had learned that he was to be deployed to Afghanistan with the National Guard. His *58brother told him that he had been sleepwalking and that he should leave. He agreed.
Throughout the interview, the detective employed a number of common investigative techniques in an effort to cause Walter to change his account. She challenged him on whether M. was lying. She repeatedly asked him to explain why M.
*183would suddenly make these allegations, and she disparaged his explanations. On multiple occasions, she expressed the opinion that he had sexually abused M. She also expressed her disbelief in his denial of culpability. She accused him of dishonesty when he omitted a detail of his conversation with Stepfather, to which she had listened. Her techniques, however, had little discernible impact on Walter's account.
On cross-examination, the detective said that M.'s parents had told her that they had changed the locks because M. had said that Walter was sleepwalking. The detective also said that Stepfather had told her that Walter sleepwalked when he was a child.7
In addition, the detective testified that M. had told her that, on the last occasion on which Walter tried to get into her room, she had screamed and awakened her mother, who had pushed Walter away. She testified that Stepfather had said that M. had said Walter had choked her, that that allegation prompted her to reinterview M., and that M. denied that Walter had choked her. Finally, she testified that M.'s parents had tried to set up an appointment with a psychologist because, they said, M. has been known to lie a lot.
Motion for Judgment of Acquittal
Following the State's case-in-chief, Walter moved for judgment of acquittal on all counts. The court granted the motion in part: it dismissed one of two counts charging second-degree rape, because M. had testified that only one rape occurred; and it dismissed a count alleging that he committed a third-degree sexual offense by digitally penetrating M.'s vagina, because M. had not testified about any such offense. The court denied Walter's motion for the offenses about which M. had testified: one instance of second-degree rape, a third-degree sexual offense, and sexual abuse of a minor.
*184The Verdict
After deliberating over the course of two days, the jury found Walter guilty of child sexual-abuse, but acquitted him of the other charges. Walter took this timely appeal.
QUESTIONS PRESENTED
Walter presents three questions for review:
1. Did the trial court err in admitting the recording of the police interview without redaction of inadmissible evidence?
2. Did the court err in permitting a State's witness to testify as an expert regarding delayed disclosure?
3. Did the court err in admitting lay opinion testimony?
For the reasons stated below, we conclude that the court erred in admitting the recorded interview without redacting the portions in which the detective questioned Walter's veracity or purported to express an opinion as to his guilt or untruthfulness. We also conclude that the record does not disclose a reliable methodology underlying Ms. Lemon's opinions, but that the State may attempt to address that deficiency on remand. We see no error on the final issue.
*59I. Expressions of Disbelief
The Court of Appeals has repeatedly held that in a criminal trial a court may not permit a witness to express an opinion about another person's credibility. See , e.g. , Fallin v. State , 460 Md. 130, 160, 188 A.3d 988 (2018) ; Hutton v. State , 339 Md. 480, 503, 663 A.2d 1289 (1995) ; Bohnert v. State , 312 Md. 266, 276-79, 539 A.2d 657 (1988).
For example, in a prosecution for child sexual abuse, the trial court allowed a social worker to give her expert opinion that the child was a victim of sexual abuse. Bohnert v. State , 312 Md. at 271, 539 A.2d 657. The social worker had based that opinion largely, if not entirely, on the child's statements. See id. The Court of Appeals reversed the conviction, because the social worker's testimony "was tantamount to a declaration by her that the child was telling the truth and that Bohnert *185was lying." Id. at 278-79, 539 A.2d 657. The Court reasoned that "the opinion was inadmissible as a matter of law because it invaded the province of the jury in two ways." Id. at 279, 539 A.2d 657. "It encroached on the jury's function to judge the credibility of the witnesses and weigh their testimony and on the jury's function to resolve contested facts." Id.
In another prosecution for child sexual abuse, the circuit court permitted the State to elicit expert testimony from a social worker that she assessed the credibility of her patients by considering the "consistency" of their statements. Hutton v. State , 339 Md. at 488, 663 A.2d 1289. The court also permitted the State to elicit expert testimony from a psychologist that the child's symptoms were "not in any way faked." Id. at 490, 663 A.2d 1289. In reversing the conviction, the Court of Appeals held that both experts commented, impermissibly, on the victim's credibility and that the psychologist opined, impermissibly, that the victim had been sexually abused. Id. at 504-05, 663 A.2d 1289. Both experts had testified about how they evaluated credibility. Id. at 505, 663 A.2d 1289. The social worker had "indicate[d] her opinion of the victim's consistency and, indirectly, her truthfulness." Id. The psychologist "necessarily stated her opinion on the victim's credibility." Id. The opinion testimony "was inadmissible, and it was error for the trial court to have allowed it." Id.
Most recently, in another prosecution for child sexual abuse, the circuit court permitted the State to repeatedly ask a psychotherapist whether she had observed signs of fabrication in the victim. Fallin v. State , 460 Md. at 145-48, 188 A.3d 988. In a 5-2 decision, a majority of the Court of Appeals concluded that the psychotherapist's testimony "appear[ed] to be indistinguishable" from the testimony that the Court had found impermissible in Bohnert and Hutton . Id. at 157, 188 A.3d 988. "The inevitable conclusion from that testimony," the majority reasoned, was that the victim's allegations "were not 'incorrect' - i.e., true." Id. Accordingly, the Court held that the testimony "impermissibly intruded on the responsibility of the jury to assess the credibility of witnesses." Id. at 132, 188 A.3d 988.
*186Just as a witness may not testify that another witness is telling the truth, a prosecutor may not ask a defendant whether other witnesses are lying. Hunter v. State , 397 Md. 580, 595-96, 919 A.2d 63 (2007). Those questions are "impermissible as a matter of law," because they encroach on the province of the jury by asking the defendant to judge the credibility of the witnesses and weigh their testimony. Id. at 595, 919 A.2d 63. Furthermore, they are "overly argumentative"; they create the risk that the jury may convict the defendant unless it concludes that the other *60witnesses are lying; they are unfair because they create a false dilemma that excludes the possibility that the other witnesses are merely mistaken; and they require a defendant to risk alienating the jury by accusing other witnesses of perjury. Id. at 595-96, 919 A.2d 63.
Walter invokes this line of cases to argue that the circuit court erred in allowing the jury to hear a number of the questions that the detectives posed to him at the recorded interview. Among others, he focuses on questions in which the detective asked whether M. was lying and whether she is "known to be [a] liar"; questions in which the detectives asked for an explanation for why M. was suddenly accusing him of abusing her and implicitly asked him to posit a motive for M. to lie;8 and questions in which the detective voiced her disbelief in what he had said and accused him of committing a crime.9 He equates the detective's questions with impermissible *187"was he lying questions" or with inadmissible expert testimony that someone was (or was not) telling the truth.
Although some out-of-state cases support Walter's argument,10 he pushes the Maryland cases farther than they can go. Cases like Bohnert , Hutton , and Fallin concern what an expert witness can say at a trial. Cases like Hunter concern what a prosecutor can ask a defendant if he or she takes the stand to testify at trial. None of those cases concern whether a jury may or may not hear questions posed by an investigator to a suspect during an interview or interrogation.
But this does not mean that the detective's questions are unproblematic: on two occasions, Maryland courts have reversed criminal convictions when a jury was exposed to an investigator's expressions of disbelief in a suspect's statements or opinions as to the suspect's untruthfulness.
In Crawford v. State , 285 Md. 431, 453, 404 A.2d 244 (1979), the Court of Appeals held that a defendant had been deprived of her due process right to a fair and impartial trial, because the jury had heard two lengthy interrogations, in which the detectives tried to persuade her to recant by *61repeatedly expressing disbelief in her claim of self-defense, asserting that she was not telling the truth, telling her that they knew how she had killed the victim, and asking her to explain why other *188witnesses (including her daughter) were lying. Id. at 439-50, 404 A.2d 244. The Court wrote:
The credibility of the accused was all important in the determination by the jury of the validity of her claim throughout the interrogations that she killed in self-defense. There is no doubt that the challenged comments of the police which were heard by the jury, whether in the form of questions, assertions of disbelief, opinions (not as expert witnesses), argument, recounting of what others were purported to have said contrary to the version of the accused, hearsay, or otherwise, tended to seriously prejudice the defense. We think that they did so improperly in the circumstances. Defense counsel made abundantly clear, time and time again, that he objected to the procedure under which the challenged matters went before the jury and to the substance of those matters.
Id. at 451.
More recently, in Casey v. State , 124 Md. App. 331, 339, 722 A.2d 385 (1999), this Court reversed a criminal conviction in part because the jury had heard a recording of an interrogation in which the investigators expressed disbelief in the defendant's account by telling him, "we know that's not true," and "we know different." Id. at 337-38, 722 A.2d 385. In his opinion for this Court, Chief Judge Joseph F. Murphy, Jr., cited Crawford for the proposition that it is "well settled that the investigating officers' opinions on the truthfulness of an accused's statements are inadmissible under Maryland Rule 5-401," which defines "relevant evidence." Id. at 339, 722 A.2d 385.11
*189On a number of occasions in the recorded interview in this case, the detective expressed her disbelief in Walter's steadfast denial of culpability. When Walter told the detective that he would never touch M. "in that manner," she replied, "[I]f I had a dime for every time I heard someone say that." In response to Walter's statement that he had been asked to leave his brother's house because he was sleepwalking, the detective said, "I don't think you sleepwalked." She promptly added: "I think you went into her room and you touched her." When Walter flatly denied that he had touched M., the detective immediately responded by suggesting that he had "[m]aybe not had the full sex, but maybe just had a feel up and touched her." When Walter rejected the detective's improbable suggestion that he might have had sexual intercourse with M. while he was sleepwalking, she volunteered: "[T]hat's why I think the touching happened, maybe not the sex." On another occasion, she bluntly accused Walter of turning M. around and having sex with her. At the end of the interview, the detective said that "somewhere between" M.'s account and Walter's "is the truth," by which she appears to have implied, again, that Walter touched M., but did not have sexual intercourse with her. Finally, on multiple occasions *62throughout the interview, the detective expressed skepticism about Walter's denials when she commented upon his inability to explain why M. would suddenly accuse him of sexually abusing her.
The expressions of disbelief were a perfectly legitimate investigative tactic to induce Walter either to confess or to change his account and to introduce inconsistencies that the detective could exploit in further questioning. Unlike the expressions of disbelief in Crawford , the objectionable comments in this case were not so pervasive as to deprive Walter of his due process right to a fair trial. Under our decision in Casey , however, the detective's expressions of disbelief were irrelevant and inadmissible. For that reason alone, we are required to reverse the conviction.
But even if we assumed that the detective's comments might have some relevance in providing context for Walter's responses, the fact is that few, if any, of Walter's responses were *190more than minimally probative in the State's case. The questions did not impel Walter to inculpate himself or to alter his account. To the contrary, Walter's account remained largely the same throughout the interview. Consequently, the detective's expressions of disbelief had little effect other than to project an aura of official skepticism over Walter's declaration of his innocence. The aura of skepticism became even more intense as a result of the detective's repeated assertions of her expertise in child sexual-abuse cases.
To make matters worse, the detective's questions and comments were often unnecessary to provide any context for Walter's answers, which is the only ostensible reason for allowing the jury to hear what the detective said. To understand that Walter repeatedly and consistently denied that he had touched M. in an inappropriate way or that he had had sexual intercourse with her, the jury did not need to hear the detective's accusations and expressions of disbelief. Because the detective was going to testify at trial anyway, the State could have asked her what Walter told her in the interview. As Walter argues, many of the detective's statements in the interview supplied "commentary not context." For that reason, any minimal probative value in the challenged statements was substantially outweighed by the considerable danger of unfair prejudice under Md. Rule 5-403.12
*191If Walter's interview was to be played to the jury, the court should have redacted the detective's commentary and *63expressions of disbelief. Alternatively, the court could have permitted the detective to tell the jury about what Walter said during the interview, without playing the recording itself. In the circumstances of this case, where the detective's comments had essentially no impact on Walter's steadfast denial of culpability, the court erred in allowing the State to play the recording of the interview without redacting the expressions of disbelief.13
Because we have concluded that the circuit court erred, we must reverse the conviction unless the error was harmless beyond a reasonable doubt. Dionas v. State , 436 Md. 97, 108, 80 A.3d 1058 (2013). "[H]armless error review 'is the standard of review most favorable to the defendant short of an automatic reversal.' " Id. at 109, 80 A.3d 1058 (quoting Bellamy v. State , 403 Md. 308, 333, 941 A.2d 1107 (2008) ). An appellate court undertaking harmless-error analysis must " 'be satisfied that there is no reasonable possibility that the evidence *192complained of - whether erroneously admitted or excluded - may have contributed to the rendition of the guilty verdict.' " Dove v. State , 415 Md. 727, 743, 4 A.3d 976 (2010) (quoting Dorsey v. State , 276 Md. 638, 659, 350 A.2d 665 (1976) ); accord Perez v. State , 420 Md. 57, 66, 21 A.3d 1048 (2011). "[W]here credibility is an issue and, thus, the jury's assessment of who is telling the truth is critical, an error affecting the jury's ability to assess a witness's credibility is not harmless error." Dionas v. State , 436 Md. at 110, 80 A.3d 1058.
On the record in this case, the State has not discharged its burden (see Perez v. State , 420 Md. at 66, 21 A.3d 1048 ) of showing that the error was harmless beyond a reasonable doubt and that it did not influence the outcome of the case. See Snyder v. State , 104 Md. App. 533, 554, 657 A.2d 342 (1995) (holding that trial court's error could not be considered harmless where court admitted a detective's testimony that "in essence, put into evidence the detective's disbelief of [the defendant's] statement regarding his activities, and the events surrounding [the crime], as well as the detective's opinion as to inconsistencies in [the defendant's] statements").
This was not an easy case. There was no physical evidence and no confession. The prosecution depended largely upon M.'s credibility, which was subject to question, not least because of the discrepancies in her several accounts. The jury, however, did not have to weigh M.'s credibility in isolation from other evidence: in deciding whether to believe M., the jury could assess Walter's demeanor in the recorded interview and the credibility of his denials. We cannot rule out the possibility that the detective's commentary on Walter's credibility *64had some effect on the jury's assessment of him.
Furthermore, there is some internal tension in the jury's verdict. The jury convicted Walter of sexual abuse of a minor, which "includes" the commission of certain underlying sexual offenses (see Md. Code (2002, 2012 Repl. Vol.), § 3-602 of the Criminal Law Article ); yet the jury acquitted Walter of the only sexual offenses that he was charged with committing. Although a conviction for child sexual-abuse is "not necessarily *193inconsistent" with an acquittal on one of the underlying offenses (see Tate v. State , 182 Md. App. 114, 129, 957 A.2d 640 (2008) ), it takes some effort to imagine how the jury could have found Walter guilty of sexual abuse, but not guilty of the only specific conduct that was alleged to have constituted the abuse. The internal tension in the jury's verdict suggests the possibility of a compromise among jurors who had real doubts about Walter's culpability. See Dionas v. State , 436 Md. at 112, 80 A.3d 1058 (citing Brooks v. United States , 367 A.2d 1297, 1310 (D.C. 1976) ).
In reversing Walter's conviction and remanding the case for a new trial, we do not mean to suggest that investigators are prohibited from expressing their disbelief in a suspect's account during an interview or interrogation, or that they may not ask a suspect whether a victim or witness is lying. Investigators can and should continue to employ lawful and effective investigative techniques. We hold only that if the State intends to play portions of a recorded interview in which the investigators directly or indirectly express their disbelief in the suspect's statements or their opinion about the suspect's guilt, the court must balance the probative value (if any) of the investigator's comments against their prejudicial effect. In general, where the investigators' comments do not induce the suspect to alter his account or to inculpate himself, a court should prohibit the State from playing those portions of the interview.14
II. Expert Testimony
Over Walter's objections, the circuit court permitted a social worker, Erin Lemon, to offer expert testimony to the effect *194that delayed reporting of child sexual abuse "is very common," especially among children who have been abused by a family member. Walter challenges both the appropriateness of Ms. Lemon's testimony and the sufficiency of the factual basis for it. He also challenges its relevance, and he asserts that its probative value was substantially outweighed by the danger of unfair prejudice.
We see no error or abuse of discretion in the court's decisions about the appropriateness of the testimony, the relevance of the testimony, or the balancing of probative value against the danger of unfair prejudice. In our view, however, the record does not disclose an adequate factual basis for the opinion, because it is unclear whether Ms. Lemon employed a reliable methodology in reaching it. The State may endeavor to correct the deficiency on remand.
A. Appropriateness and Factual Basis
Md. Rule 5-702 governs the admissibility of expert testimony:
*65Expert testimony may be admitted, in the form of an opinion or otherwise, if the court determines that the testimony will assist the trier of fact to understand the evidence or to determine a fact in issue. In making that determination, the court shall determine (1) whether the witness is qualified as an expert by knowledge, skill, experience, training, or education, (2) the appropriateness of the expert testimony on the particular subject, and (3) whether a sufficient factual basis exists to support the expert testimony.
"[T]he trial court must make the determinations set forth in the rule before ruling on the admissibility of expert testimony, but its decision as to whether to admit the testimony is discretionary." Sissoko v. State , 236 Md. App. 676, 712, 182 A.3d 874, cert. denied , 460 Md. 1, 188 A.3d 917 (2018). "For that reason, we review that decision for abuse of discretion[.]" Id. (citing Rollins v. State , 392 Md. 455, 500, 897 A.2d 821 (2006) ). "[A] court's 'action in admitting or excluding *195such testimony will seldom constitute a ground for reversal.' " Id. (quoting Bryant v. State , 393 Md. 196, 203, 900 A.2d 227 (2006) ).
Walter challenges both the appropriateness of Ms. Lemon's testimony and the sufficiency of the factual basis for it.15
To determine the appropriateness of expert testimony on a particular subject, a court should ask "whether the trier of fact will receive appreciable help from the expert testimony in order to understand the evidence or to determine a fact in issue." Sippio v. State , 350 Md. 633, 649, 714 A.2d 864 (1998).
Walter argues that Ms. Lemon's testimony "was not of 'appreciable help' to the jury in understanding the evidence," because, he says, her testimony, that "delayed disclosure occurs in cases of actual abuse and in cases of fabrication," does not aid the trier of fact in understanding M.'s disclosure. The State responds that Ms. Lemon's testimony explains the behavior of delayed reporting, which may seem inconsistent with having been sexually abused. The trial court found Ms. Lemon's testimony to be appropriate because "lay persons and jurors are not aware of the factors that may go into delayed reporting."
In our judgment, this issue is controlled by the decision in Yount v. State , 99 Md. App. 207, 636 A.2d 50 (1994). In that case the trial court allowed expert testimony from a therapist who had a master's degree in psychology, approximately 150 hours of training on child sexual-abuse, and professional certifications as a mental health counselor. Id. at 210, 636 A.2d 50.
*196The therapist testified that, "as a general phenomenon," victims of child sexual-abuse frequently recant their initial allegations of abuse. Id. In addition, she explained that it was "very common" for a child to retract the recantation "once the child's support systems were in place." Id. at 211, 636 A.2d 50.16
*66In response to an appellate challenge to the appropriateness of the expert testimony, this Court in Yount held that the psychological phenomenon of wavering on an accusation is "not part of the common currency of lay experience[,]" and thus that knowledge of such a phenomenon would be of appreciable help to the fact finder. Id. at 211-12, 636 A.2d 50. Without expert guidance, the factfinder might "easily have fallen into the untutored lay[person]'s error of dismissing as noncredible testimony that, in the arcane context of sexual child abuse, should not be so readily dismissed." Id. at 212, 636 A.2d 50.
Similarly, in this case, Ms. Lemon provided general testimony about the phenomenon of delayed reporting and why some children may delay in reporting the abuse. She provided the jury with a balanced explanation of delayed reporting, supporting her testimony with a number of factors that could result in delayed disclosure, including factors other than sexual abuse. In these circumstances, the court could reasonably conclude that Ms. Lemon's testimony provided appreciable help to the factfinder about a phenomenon that was not within a layperson's knowledge; her testimony, therefore, met the requirement of appropriateness in Rule 5-702.
Walter, however, does not just challenge the appropriateness of Ms. Lemon's testimony; he also challenges the sufficiency of the factual basis for it. The sufficiency of the factual basis "include[s] two subfactors: an adequate supply of *197data and a reliable methodology." Rochkind v. Stevenson , 454 Md. 277, 286, 164 A.3d 254 (2017) (citing Roy v. Dackman , 445 Md. 23, 42-43, 124 A.3d 169 (2015) ); accord Savage v. State , 455 Md. 138, 183, 166 A.3d 183 (2017).
Ms. Lemon undoubtedly had an adequate supply of data from her personal experiences, but the reliability of her methodology is another question. From the record before us, it is unclear how Ms. Lemon concluded that victims of child sexual abuse often, frequently, or commonly delay in reporting the abuse. In particular, it is unclear how she determined that a delayed report originated with a bona fide victim as opposed to someone who had fabricated a report or had a false memory of abuse, which, she recognized, sometimes occurs. Ms. Lemon kept no statistics and could point to no peer-reviewed studies to support her conclusion, so she appears to have based her opinion on only an extrapolation from her own experiences. In evaluating those experiences, did she do anything to distinguish true or reliable claims from false or unreliable claims? For example, did she assume that a person was a victim of sexual abuse only if an abuser has been convicted of sexual abuse, or if the abuser has admitted to sexual abuse, or if there is some corroborating evidence of sexual abuse? Did she rely on her own, subjective evaluation of the validity of the claim of abuse? Or did she draw the conclusion from a conflation of all of the claims that she had heard, without distinguishing the true from the false or the reliable from the disproven? We simply do not know.
Tellingly, the State's brief defends the supply of data for Ms. Lemon's opinion, but is silent as to the reliability of her methodology. Consequently, we must conclude that the record does not disclose a sufficient factual basis for the opinions. On remand, the State is free to develop the record to establish the reliability of the *67methodology and thus the sufficiency of the factual basis for the opinion.
B. Relevance and Unfair Prejudice
Walter argues that Ms. Lemon's testimony was irrelevant, because he affirmatively said that he would not use the *198delayed disclosure as a basis to attack M.'s credibility. Alternatively, he argues that the probative value of Ms. Lemon's testimony was substantially outweighed by the danger of unfair prejudice, because, he says, its only purpose was to bolster M.'s credibility. We are unpersuaded.
Evidence is relevant if it tends to "make the existence of any fact that is of consequence to the determination of the action more probable or less probable than it would be without the evidence." Md. Rule 5-401. A court may admit relevant evidence, but it has no discretion to admit evidence that is irrelevant. Smith v. State , 218 Md. App. 689, 704, 98 A.3d 444 (2014) (citing Md. Rule 5-402 ). A ruling that evidence is legally relevant is a conclusion of law, which we review de novo. See id.
Even if evidence is relevant, however, a court may exclude it "if its probative value is substantially outweighed by the danger of unfair prejudice, confusion of the issues, or misleading the jury, or by considerations of undue delay, waste of time, or needless presentation of cumulative evidence." Md. Rule 5-403. We review that decision for abuse of discretion. See , e.g. , Carter v. State , 374 Md. 693, 705, 824 A.2d 123 (2003).
Notwithstanding Walter's agreement not to use the issue of delayed disclosure to attack M.'s credibility, Ms. Lemon's testimony was relevant, because it would "assist the trier of fact to understand the evidence or to determine a fact in issue[.]" State v. Smullen , 380 Md. 233, 268-69, 844 A.2d 429 (2004). Even if the defense did not raise the issue of delayed disclosure, the State was entitled to anticipate that a juror might cite the delay as a reason not to believe M.'s testimony. In a case in which credibility was key, and in which the defense asserted that M.'s account was "riddled with inconsistencies, lies, exaggerations, explanations that defy logic, [and] actions and behavior that violated [ ] common sense," Ms. Lemon's testimony could assist the jury in understanding why a young victim might not report the abuse until more than three years after it occurred.
*199Furthermore, the court did not abuse its discretion in concluding that the probative value of Ms. Lemon's testimony was not substantially outweighed by the danger of unfair prejudice. Here, the decision in Yount is decisive again. In Yount the expert testified that victims of child sexual-abuse frequently recant their initial accusations and later recant their recantations, but she did not testify on the ultimate issue of whether the alleged victim in that case had been sexually abused. In other words, the testimony "advised the jury as to the existence of a psychological phenomenon that would explain the child sexual abuse victim's wavering or vacillation." Yount v. State , 99 Md. App. at 218-19, 636 A.2d 50. "Beyond that," however, "it did nothing to indicate that the victim's version of events rather than the appellant's version of events should be believed." Id. at 219, 636 A.2d 50.
The same is true in this case: Ms. Lemon's testimony, which may have been modeled on the testimony in Yount , told the jury why a victim of child sexual-abuse might delay in reporting the abuse, but it did nothing else to indicate that the jury should prefer M.'s version over Walter's. Under Yount , the danger of unfair prejudice did not substantially outweigh the probative value of the expert's testimony.
*68III. Lay Opinion Testimony
Stepfather testified that before asking Walter to move out of the house in 2012, he called their father to help him. Stepfather and his father sat Walter in a lawn chair and told him that he had scared M. and that he needed to leave. Stepfather stated that during the conversation Walter had his "eyes wide open," with his hands on the chair, "just sitting back," "like he's trapped." After the conversation, Walter immediately packed his belongings and left the house.
Walter argues that the court erred by admitting lay opinion testimony when, over objection, Stepfather was permitted to testify that Walter looked "trapped" when he was confronted and asked to leave the house. We disagree.
*200"The admissibility of evidence ordinarily is left to the sound discretion of the trial court." Moreland v. State , 207 Md. App. 563, 568, 53 A.3d 449 (2012). "We will not disturb a trial court's evidentiary ruling unless 'the evidence is plainly inadmissible under a specific rule or principle of law or there is a clear showing of an abuse of discretion.' " Id. at 568-69, 53 A.3d 449 (quoting Decker v. State , 408 Md. 631, 649, 971 A.2d 268 (2009) ) (further quotation marks omitted). "A court's decision is an abuse of discretion when it is 'well removed from any center mark imagined by the reviewing court and beyond the fringe of what that court deems minimally acceptable.' " Id. at 569, 53 A.3d 449 (quoting Gray v. State , 388 Md. 366, 383, 879 A.2d 1064 (2005) ) (further quotation marks omitted).
Md. Rule 5-701 governs lay opinion testimony. It provides:
If the witness is not testifying as an expert, the witness's testimony in the form of opinions or inferences is limited to those opinions or inferences which are (1) rationally based on the perception of the witness and (2) helpful to a clear understanding of the witness's testimony or the determination of a fact in issue.
Rule 5-701"reflects the modern trend away from the traditional common law tenet that a lay witness may not express an opinion." Robinson v. State , 348 Md. 104, 118, 702 A.2d 741 (1997).
"Lay opinion testimony is testimony that is rationally based on the perceptions of the witness." Ragland v. State , 385 Md. 706, 717, 870 A.2d 609 (2005). " 'The prototypical example of [proper lay opinion testimony] relates to the appearance of persons or things, identity, the manner of conduct , competency of a person, degrees of light or darkness, sound, size, weight, distance, and an endless number of items that cannot be described factually in words apart from inferences.' " Id. at 718, 870 A.2d 609 (quoting Asplundh Mfg. Div. v. Benton Harbor Eng'g , 57 F.3d 1190, 1196 (3d Cir. 1995) ) (emphasis added). "A trial court should, within the sound exercise of its discretion, admit lay opinion testimony if such testimony is derived from first-hand knowledge; is rationally *201connected to the underlying facts; is helpful to the trier of fact; and is not barred by any other rule of evidence." Robinson v. State , 348 Md. at 118, 702 A.2d 741. For instance, a trial court may admit lay opinion testimony that a defendant appeared to be nervous when he was approached by police officers. See Jones v. State , 132 Md. App. 657, 679-80, 753 A.2d 587 (2000).
Lay opinion testimony is permissible "where it is impossible, difficult, or inefficient to verbalize or communicate the underlying data observed by the witness." Id. at 119, 702 A.2d 741. Stepfather's testimony fits that description.
Stepfather's testimony, that Walter "was putting his hands on the chair ... just *69sitting back like this, like he's trapped," was his interpretation of how he perceived his brother's conduct at the time of the confrontation. Allowing Stepfather to say that Walter looked "trapped" was much more efficient than requiring him to identify each of the minute observations that led him to his conclusion.
Stepfather's testimony did not undercut the role of the jury as factfinder. The jurors were free to assign as much or as little weight as they saw fit to Stepfather's assessment of his brother's demeanor. We see no error or abuse of discretion in the court's decision to admit Stepfather's testimony.
JUDGMENT OF THE CIRCUIT COURT FOR ANNE ARUNDEL COUNTY REVERSED. CASE REMANDED FOR FURTHER PROCEEDINGS CONSISTENT WITH THIS OPINION. COSTS TO BE PAID BY ANNE ARUNDEL COUNTY.