FLYNN, J.
Defendant appeals a judgment of conviction for sexual abuse in the first degree, ORS 163.427. Defendant assigns error to the trial court’s pretrial exclusion of expert testimony that defendant offered to support his contention that his admissions to touching the victim were a product of an adjustment disorder and were misunderstood as confessions of abuse. We conclude that the trial court erred in excluding that testimony regarding diagnoses of defendant’s mental health issues, and we further conclude that the error was not harmless. We therefore reverse and remand.
I. BACKGROUND
Defendant and his wife, S, were married for eight years and have two children, M and L. Defendant was indicted in April 2011 for one count of sexual abuse in the first degree based on an allegation of sexual contact with M between March 2009 and March 2010. Because defendant argues that the excluded expert testimony would have helped to explain statements and behavior to which the state pointed as admissions of guilt, we describe those statements and behavior in detail. M was four years old in November 2009, when S returned home from work one morning and found defendant extremely upset. As S later described the discussion, defendant told her that she would never forgive him, handed over his wedding ring, and then told her that he had touched M over her diaper.1 S testified that defendant also said that he had been thinking about touching the children for some time and that he was worried about “turning into his family.” Although S did not understand the family comment at the time, defendant explained at trial that he “had a couple of * * * cousins that have been actually sexually molested, and [he] was at the place when it happened” and that he had “heard rumors” about his mother “possibly being molested by somebody.” According to S, defendant said that “he belonged in prison and to call the police right now.” S did not call the police because defendant agreed to leave [4]*4the house and because defendant told her that M had not awakened and did not know about the touching.
Defendant acknowledged that he was very emotional when S came home and that he had thoughts about “my cousin and my mom being molested as a kid, and I felt that that was pretty disgustingU” He testified, however, that he told S he only accidentally touched M’s diaper while pulling up her covers and that S “freaked out.” He insisted that he “really didn’t think much of it other than the fact it was an accident” and that he took off his ring and left the house because of S’s reaction. He described his thoughts as, “[Y]ou know, you’re freaking out at me, I’m done, I’m going to walk away here.”
S agreed to stay married to defendant for a short time so that he could seek counseling using her health insurance. S insisted, however, that she “needed some sort of proof that says that [defendant] did this so that * * * he couldn’t come back later and say that he wanted the kids.” S testified that she wrote the following statement for defendant to sign:
“I, [defendant], am admitting to touching my oldest daughter, [M], in an inappropriate way. I went in her room early in the morning and was tucking her into bed when I ran my hand over her diaper, over her pubic area. I then walked out and went to bed. My daughter did not wake up. I have never done this before. When my wife came home from work, I told her. She told me to leave and I agreed to that and to get counseling. I am writing this so my wife, [S,] has proof for the protection of our children.”
S testified that she read the document out loud to defendant, and then defendant read and signed it. At trial, defendant took the stand and denied that S showed him the document and claimed that his signature was forged.
S and defendant divorced in March 2010. On September 11, 2010, M told S, “Well, one time I was in bed with daddy and — and he put his hand down my diaper.” M moved her hand over her pubic area to show S how defendant had touched her and said that defendant’s hand was on “her private.” M said that defendant did not take his hand out, so M pulled his hand out of her diaper. M told S that she [5]*5did not tell about the incident sooner because she was afraid that defendant “would have to go away.”
A few days later, before S called the police, she had M repeat the story. M’s story was the same except for an added detail about defendant saying, “I’m sorry.” M described the incident similarly in a recorded interview with a medical examiner at CARES Northwest on September 20, 2010, and in testimony at trial. The CARES medical examiner did not find any physical signs of abuse, but testified that that was consistent with an allegation of touching the outside of the vagina.
Before the disclosure by M, defendant attended counseling with Dr. Callum, a licensed psychologist, for 15 sessions between December 2009 and July 2010. Defendant discussed with Callum issues of anxiety and marital difficulties. He told Callum that he was concerned about “thoughts” of touching his children but assured Callum that he had “never” acted on the thoughts. Defendant later insisted that he never had inappropriate thoughts but that he and S agreed that he would tell Callum otherwise in order to avoid telling her about the touching incident, which defendant feared Callum might have to report to the authorities. Callum was not aware of the signed statement during the time that she was providing therapy.
Defendant was arrested on April 3, 2011. Nearly a year later, defendant was in the Medical Observation Unit of the Washington County Jail when he approached two deputies and asked, “What happens if I confess right now?” A few minutes later, defendant approached them again and said, “I did it. I confess.” A deputy told defendant that the matter was between him and the courts. Defendant continued to pace the unit, approached the deputies again, and said, “Okay. I touched my daughter. I admit [it]. I’m a jerk.”
Before trial, it became apparent to the state that defendant intended to call Callum as one of his witnesses to testify about her clinical observations and opinions about defendant. The state filed a motion in limine to exclude her testimony, arguing, among other things, that the evidence was inadmissible under OEC 702, which provides that, “[i]f scientific, technical or other specialized knowledge will assist [6]*6the trier of fact to understand the evidence or to determine a fact in issue, a witness qualified as an expert by knowledge, skill, experience, training or education may testify thereto in the form of an opinion or otherwise.” (Emphasis added.) Relying on our opinion in State v. Nichols, 252 Or App 114, 118, 284 P3d 1246 (2012), rev den, 353 Or 428 (2013), the state argued that “defendant will not be able [to] establish a sufficient nexus between [his] actions/behaviors and * * * Callum’s observations, opinions, discussions, recommendations, or beliefs surrounding her contacts with the defendant.” The state contended that “Callum has not even provided a formal diagnosis, and if such a diagnosis were to be provided, there is not a satisfactory link that can be demonstrated between the diagnosis and the material events that occurred in this case”; thus, her “testimony is not assistive to the trier of fact and should be excluded.”
The court held a pretrial hearing at which defendant elicited extensive testimony from Callum as an offer of proof. On direct examination during that proffer, Callum testified that defendant
“had very low self-esteem. He was also plagued by fears and many of the cognitions of his childhood, which kept on just ruminating [in] his head, and those were the fears and negative thoughts that he was always having. And which continuously made him wonder if he was okay, what was wrong with him.”
Callum testified that, based on her clinical observations, defendant’s “level of rumination” could “definitely lead him to become obsessive” and, at times, “be unhealthy.”
Later in her direct examination, Callum returned to the subject of defendant’s mental health, and she was asked whether he had “adequate coping skills and self-esteem to deal with personal issues.” She testified, “No. No. A lot of the times he — he was very fearful, and he let * * * the rumination of his thoughts plague. He could not let those things go.” On cross-examination, she further explained the nature of defendant’s adjustment disorder and mental health issues:
“A. [CALLUM]: * * * I diagnosed him as adjustment disorder with depressed mood. And I also diagnosed him with a trauma related to childhood physical and emotional [7]*7abuse. And then the AXIS IV would be (indiscernible) experiencing marital difficulties or extramarital affairs.
“Q. [PROSECUTOR]: What is adjustment disorder?
“A. [CALLUM]: Adjustment disorder is when they have difficulty in [being] able to cope with the situation or where they have — just not being able to — to put everything together and they are letting it affect them. And you — and there’s several — you can have adjustment disorder not otherwise specified, you can have adjustment disorder with anxiety and depression. You can have adjustment disorder with anxiety alone, you can have it with depression alone. You can also have it — adjustment disorder with a variety of emotions. And that is a very typical diagnosis when people are having difficulty in bringing it all together.”2
Also on cross-examination, Callum addressed the relationship between defendant’s mental health diagnoses and his statement to the jail deputies:
“Q. [PROSECUTOR]: Were you aware that [defendant] confessed in March of 2012 to touching his daughter, to two jail deputies, without them asking any questions?
“A. [CALLUM]: I heard about that, yes.
“Q. [PROSECUTOR]: Okay. And that’s factoring into your equation as well, when you say to a reasonable degree of medical certainty, you can testify under oath that he did not molest any of his children or touch them in an inappropriate way?
“A. [CALLUM]: In regards to that confession-—
“Q. [PROSECUTOR]: I’m asking you whether you’re factoring that in.
“A. [CALLUM]: And that’s what I’m going to say. In regards to that confession, I would factor my knowledge [8]*8about him, about his adjustment, you know his high degree of distress and so that — that would be where with my medical certainty would be he was very distressed. And whether he’s confessing to something out of distress, that’s within the realm of possibility.
* * * %
“Q. [PROSECUTOR]: And all that [a signed confession, disclosure of thoughts of sexually touching children, and the jailhouse admissions], even knowing that, today you’re testifying under oath, that you believe to a reasonable degree of medical certainty that he did not abuse his children?
‘A. [CALLUM]: That’s correct.”
Following the offer of proof, the state argued, as it had in its written motion, that “there was no description by *** Callum, as to how this adjustment disorder would have affected [defendant’s] likelihood of confessing, making false confessions, writing false confessions, anything.” The state further contended that “there’s been no evidence that would connect his mental health issues, or whatever he’s going through, his adjustment disorder, to any sort of other behavior, of how that would affect his behavior throughout the course of trial — and again, that’s what * * * this witness is being proffered for.”
The trial court then expressed its view that defendant’s adjustment disorder was not “relevant” to the issues in the case. The court explained:
“I’m looking at the diagnosis of his adjustment disorder with depression, thinking well, what does that have to do with anything? I’m looking at the — he was there to be treated by — he was there to be treated by Dr. Callum, the wife wasn’t there for treatment, the child wasn’t there for treatment, she had very little contact with either one of them in a treatment phase. Certainly she is not going to be able to get — as you said yourself, she can’t give the opinion of what she — that she believes he didn’t do the act, and all that kind of stuff.
“But, you know, the fact that he suffers from adjustment disorder with depression, has really nothing to do with this case.
[9]*9“*** I’m having a hard time figuring out the rest of this, because I don’t know what the rest — you were talking about you wanted to get into his psychological profile, and I’m thinking well, how is that relevant to his not understanding what he signed or — I mean, other than that, I don’t know what it’s relevant to.”
(Emphases added.)
Defense counsel then tried to address the court’s concern about the connection between Callum’s testimony and his theory of the case. Counsel argued that defendant’s “psychological profile also applies to the statements of the guards, because of his poor coping skills and his excessive rumination and his hallucinatory behavior,” and that “this is the nexus right out of Nichols, is that people who have poor coping skills, try different strategies, and — and people make false confessions, and make false statements.”
Counsel emphasized that the theory of defense was that the defendant’s hand
“brushed the front of [M’s] diaper, and that what it was in fact, was an accidental touching, but that [defendant], because of his excessive rumination, as described by Dr. Callum, his poor life skills, his poor coping skills, his poor understanding of the way things work, I believe she described him as being very simplistic, he began ruminating over that and wondering because of his familial origin, if he was turning into the kind of person that would do something like this to his daughter.”
(Emphasis added.)
Counsel further emphasized that he was “not just coming up with this”; when defendant was evaluated by a different psychologist, Dr. Czar, defendant had denied any sexual motive in the touching and had insisted that
“what we said to Dr. Callum, which was 1 had bad thoughts,’ was something that he and his wife hatched together to get him services. That he was concerned, that because of his family origin, that he might have a subconscious motive that was in play.”3
[10]*10Defense counsel further explained that the “thrust of our case” is that, because of defendant’s “poor coping skills, excessive ruminating, which is what Dr. Callum described, adjustment disorders,” he deferred to his wife and signed the confession when she presented it, “because his overwhelming concern was, ‘I’ll do whatever you guys want me to do, I just want to see my children again. If you — if you want me to have treatment, I’ll do treatment. I just want to see my kids again.’” Counsel explained that, “under our theory, that he has poor coping skills and makes bad choices because of poor coping skills, which [Callum] testified to, that occurs in false confession cases.”
The trial court agreed with the state and ruled that Callum “cannot get into his psychological diagnosis or profile.” The court explained that the “diagnosis is [not] helpful to the jury. There’s no nexus between that diagnosis and the defense in this case.” (Emphasis added.)
At trial, the state offered evidence of defendant’s conduct, as described above: repeated admissions to S about touching M and having sexual thoughts about his children, his admissions to Callum that he had sexual thoughts about his children, and testimony from M that defendant had abused her. Defendant took the stand and testified in his own defense. The jury found defendant guilty, and he was convicted of first-degree sexual abuse.
II. ANALYSIS
On appeal, defendant argues that the trial court erred by granting the state’s motion to limit Callum’s testimony concerning defendant’s mental health issues. Defendant argues, as he did below, that Callum’s testimony about his adjustment disorder and corresponding “very, very poor coping skills” were relevant to his theory that the disorder “led him to ruminate and obsess over the sexual abuse that occurred in his family when he was a child and to consequently overreact to and misinterpret his accidental touching of M.” Defendant’s theory, he argues, is that he “made statements to [S] and the jail guards. Others interpreted those statements as admissions of touching M with a sexual purpose though defendant did not acknowledge any sexual intent.” According to defendant, expert testimony [11]*11that would explain defendant’s “ruminations and obsession” would have been helpful to the trier of fact when assessing defendant’s abnormal behavior. The state responds, again relying on Nichols, that the trial court properly excluded the evidence under OEC 702, because defendant failed to establish a sufficient nexus between Callum’s diagnosis and the behavior that defendant sought to explain.
A. Standard of Review
“We review the trial court’s order to exclude expert testimony to determine whether the court applied the correct principles of law and did not abuse its discretion.” State v. Gherasim, 329 Or 188, 198, 985 P2d 1267 (1999). Discretion in the context of evidentiary rulings “refers to the authority of a trial court to choose among several legally correct outcomes.” State v. Rogers, 330 Or 282, 312, 4 P3d 1261 (2000). Thus,
“we first must review evidentiary rulings without deference to determine whether proper principles of law were applied correctly. * * * Only if we determine that application of the correct legal principles leads to more than one correct outcome do we continue to review whether the trial court abused its discretion in choosing an outcome.”
Id.
B. OEC 702
In light of that standard of review, we begin by describing the legal principles implicated by the state’s motion to exclude Callum’s testimony. We have observed that the “test under OEC 702 has been stated rather simply as ‘whether the expert’s testimony, if believed, will be of help or assistance to the jury.’” Nichols, 252 Or App at 119 (quoting State v. Stringer, 292 Or 388, 391, 639 P2d 1264 (1982)). In application, however, the test is more complex. As the court explained in State v. O’Key, 321 Or 285, 298, 899 P2d 663 (1995):
“The requirement in OEC 702 that the evidence or testimony ‘assist the trier of fact to understand the evidence or to determine a fact in issue’ is intended to serve multiple functions, such as:
[12]*12“(1) supplying general propositions which will permit inferences from data which the trier of fact would otherwise be forced to find meaningless; (2) applying general propositions to data so as to generate inferences where the complexity of the body of propositions applied, the difficulty of the application, or other factors make the expert’s conclusion probably more accurate or precise than that of the trier of fact; (3) modifying, qualifying, and refining general propositions which the trier of fact may reasonably be expected to use; and (4) adding specialized confirmation and, thus, confidence to general propositions otherwise likely to be assumed more tentatively by the trier.”
(Quoting John William Strong, Language and Logic in Expert Testimony: Limiting Expert Testimony by Restrictions of Function, Reliability, and Form, 71 Or L Rev 349, 360 (1992).)
C. State v. Nichols
To serve any of those multiple functions, the expert’s testimony must be explained in a way that will permit a trier of fact to draw reasonable inferences about pertinent issues in the case.4 We applied that principle in Nichols. Because that decision has been the focus of much of the parties’ argument throughout this case, we describe the facts of Nichols in detail.
In Nichols, the state tried the defendant on the theory that she had murdered her neighbor in the course of stealing the neighbor’s Oxycodone and other property. As [13]*13part of its case-in-chief, the state offered testimony from investigating detectives about the defendant’s behavior during an interview at her home after the murder. Toward the end of that interview, the defendant had asked whether she could go to the store to get some pain medication, and the detectives informed her that they would first like to complete their search of her home. The defendant agreed, “but then ‘sprung up from the couch’ and ‘bee-line [d] down the hall.’” 252 Or App at 116-17. The detectives followed her and watched her “remove a videotape from her closet and a small plastic bag apparently concealed in the videotape cover. Then, in plain view of the detectives, she moved in a way that indicated that she was removing a ring from her finger, and then quickly placed her hand in her front pants pocket.” Id. at 117. The detectives asked the defendant what she was doing, and she told them “nothing.” Id. She then consented to a search, and the detectives discovered a large gold and diamond ring and a plastic bag containing Oxycodone. In response to further questions, the defendant admitted that she had stolen the pills and a ring from the victim but claimed to have done so long before the date of the murder. The defendant also made inconsistent statements about her theft of the Oxycodone. Id.
To counter that testimony, the defendant sought to introduce testimony from her counselor about her mental health. The defendant contended that, “because the state was using her inappropriate responses and behavior during the investigation to imply that she was guilty, she should be allowed to introduce evidence, in the form of expert testimony, of possible other explanations for her conduct.” Id. The counselor’s testimony was “relevant to show that her behavior was not an indication of guilt but, rather, was a result of her mental illness.” Id.
The counselor’s testimony was then presented during an offer of proof outside the jury’s presence. The counselor testified that he had performed a health assessment that revealed that the defendant suffered from bipolar disorder and borderline personality disorder, had a history of substance abuse and kleptomania, and demonstrated self-destructive behavior. Id. at 117-18. He also mentioned that he had observed “manifestations of [] defendant’s [14]*14self-destructive behavior in the form of ‘picking on herself and scars * * * from the scabs healing over and then * * * she would pick on those scabs and, and reinjure herself in that fashion.’” Id.
The trial court excluded the proffered evidence as unhelpful, and we affirmed, concluding that the counselor’s testimony would have given the jury the diagnoses without “any explanation as to how those diagnoses might generally affect someone’s behavior, let alone how they might have affected [] defendant’s behavior.” Id. at 121.
In the course of our analysis, we distinguished Gherasim, 329 Or at 188, a case in which the Supreme Court concluded that the trial court should have admitted an expert opinion that “the victim likely suffered from dissociative amnesia and explained that, as a result, her capacity to remember what had occurred on the night she was assaulted was impaired.” Nichols, 252 Or App at 121. Our decision emphasized that the evidence in Nichols lacked the link between the defendant’s mental health issues and her behavior that, in Gherasim, made the diagnosis evidence helpful to the jury. Without such a link, we explained, the expert evidence would have left the jury “with diagnoses but no further explanation as to how those diagnoses might affect” the behavior at issue and, thus, “would not have assisted the jury in assessing defendant’s behavior.” Id. Again contrasting Gherasim, we explained:
“In [that case], the expert’s failure to explain his diagnosis in detail was not fatal to the admissibility of his testimony, but here defendant failed even to offer the level of explanation offered in Gherasim. That is, in Gherasim the expert explained that the victim’s condition affected her capacity to remember and to accurately recount what occurred on the night that she was assaulted. In this case, defendant’s counselor failed to offer even a rudimentary explanation that would link defendant’s behavior to her mental health issues. Accordingly, the trial court did not err in excluding the testimony of defendant’s mental health counselor.”
Nichols, 252 Or App at 121 (emphasis added).
[15]*15D. Callum’s Excluded Testimony
In this case, the trial court excluded Callum’s testimony relating to defendant’s adjustment disorder on the ground that “[t] here’s no nexus between that diagnosis and the defense in this case,” presumably drawing on our decision in Nichols. However, Nichols does not support the court’s ruling.
As just described, Nichols involved a circumstance in which the defendant sought to introduce expert testimony to explain her behavior, but then failed to elicit testimony that supplied “even a rudimentary explanation that would link [] defendant’s behavior to her mental health issues.” Consequently, the counselor’s diagnosis, standing alone, was not pertinent to any issue in the case, because a fact-finder would have been required to speculate about how the diagnosis had any bearing on the facts at issue.
By contrast, defendant’s offer of proof in this case connected defendant’s diagnosis of an adjustment disorder to defendant’s behavior that was at issue. Callum testified that a person with an adjustment disorder might “have difficulty in [being] able to cope with the situation or where they have — just not being able to — to put everything together and they are letting it affect them”; that defendant had an adjustment disorder and “very, very, poor coping skills”; that, as a result of his poor coping skills, he had a “level of rumination” that at times led to unhealthy obsession; and that he was “plagued by fears and many of the cognitions of his childhood, which kept on just ruminating [in] his head, and those were the fears and negative thoughts that he was always having” and that “continuously made him wonder if he was okay, what was wrong with him.”
On this record, unlike in Nichols, it is clear that the expert testimony, if believed, could have helped the jury assess defendant’s behavior. As defendant points out, his theory of the case — which he articulated before trial when the court ruled on the motion in limine — was that he accidentally touched his daughter on her diaper while pulling up her covers; that he “began ruminating over “that accidental touching incident and then reported it to his wife out of concern, [16]*16“because of his family origin, that he might have a subconscious motive that was in play”; and that defendant, overwhelmed by the stress of his wife’s reaction and the situation, simply agreed to sign a confession in an effort to defuse matters, without fully understanding its significance. Callum’s testimony was consistent with and would have bolstered that theory, providing a scientific basis — an adjustment disorder that made defendant susceptible to rumination and obsessive thinking, and left him with very poor coping skills — to help explain what would be a highly unusual reaction to an accidental touching. See O’Key, 321 Or at 298 (explaining that expert testimony can be helpful by adding “specialized confirmation and, thus, confidence to general propositions otherwise likely to be assumed more tentatively by the trier of fact”). Moreover, Callum’s testimony would have supported defendant’s theory that he later tried to confess to the jail guards because he had been overwhelmed by the stress of the situation. Indeed, that is precisely the connection Callum offered on cross-examination to explain why, although she was aware of defendant’s statements to the jail guards, she did not view them as a true admission of guilt.
In short, the record supplies a nexus between Callum’s testimony about defendant’s adjustment disorder and its effect on defendant. That testimony, if believed, would have helped the jury to assess defendant’s explanation for why he would have told his wife about touching M, why he would have signed the written confession, and why he would have later confessed to jail deputies, if defendant had only accidentally touched M, as he maintained. The trial court therefore erred in excluding Callum’s diagnosis testimony on the ground that it lacked a nexus to defendant’s behavior.5 See Gherasim, 329 Or at 198.
E. Harmless Error Analysis
The remaining question is whether that error was harmless. See Or Const, Art VII (Amended), § 3 (an appellate court must affirm a conviction, notwithstanding any evidentiary error, if there is little likelihood that the error [17]*17affected the verdict); State v. Davis, 336 Or 19, 32, 77 P3d 1111 (2003) (explaining that the analytical focus is “on the possible influence of the error on the verdict rendered, not whether [the reviewing court], sitting as a factfinder, would regard the evidence of guilt as substantial and compelling”).
According to the state, the exclusion of Callum’s testimony was harmless in light of the fact that defendant himself took the stand and testified in a manner that was inconsistent with some parts of the defense theory that he advanced during the hearing on the motion in limine. The state points out that, with respect to his initial statements to his wife, defendant insisted at trial that he had described the incident as an accidental touching and his wife had overreacted. With respect to the written confession, defendant testified that he had only seen the document in preparation for trial and that his signature was forged. And, with respect to his statements to the jail deputies, defendant testified at trial that he was “processing just fine” and not under any stress when he made those statements. Given that testimony, the state argues, “Callum’s proposed testimony would not have helped the jury in understanding defendant’s conduct or statements, [and] any error in excluding her testimony was necessarily harmless.”
Although defendant’s own trial testimony muddled the theory initially advanced by defense counsel,6 other [18]*18evidence indicated that defendant signed the written admission and was in distress when he confessed while in the jail Medical Observation Unit. Moreover, Callum’s testimony regarding defendant’s mental health would have been helpful to his defense even in light of defendant’s trial testimony. First, although defendant insisted that he only described an “accidental” touching to his wife, he confirmed his wife’s testimony that he was very emotional when she returned home and spontaneously told her that he had touched his daughter’s diaper. He also testified, “And, you know, that morning, and me thinking about, you know, what happened to my cousins and stuff [involving sexual abuse] that — it just — I don’t know. I guess because the state of depression I told her that, you know, it was an accident, I was depressed, and because I was crying ***.” There was also evidence that defendant had told clinical psychologist Czar that he had told his wife about the incident to “play it safe” because “‘there was a part of [defendant] that thought maybe there was more to it’” and that he was concerned that there “could have been an unconscious sexual motive.” Callum’s testimony would have allowed defense counsel to explain why defendant would, as a result of obsessive thinking and adjustment disorder, have admitted the touching to his wife and then signed the written statement she prepared for him, even if the touching were accidental.
Moreover, although defendant testified that he was “processing” fine and not under stress at the time he confessed to jailhouse deputies, he also testified, “I honestly don’t know why I would have said that I had touched my daughter.” Callum’s testimony would have allowed defense counsel to supply an alternative explanation for defendant’s seemingly incriminating behavior — even if defendant was not consciously aware that he was under stress at the time that he made the statements.
Finally, this is not a case in which we can conclude that there was “overwhelming evidence” that defendant committed the abuse regardless of whether Callum’s testimony is considered. Cf. State v. Harding, 221 Or App 294, 302, 189 P3d 1259, rev den, 345 Or 503 (2008) (concluding that error in admitting the defendant’s inculpatory statements was harmless “in- light of the overwhelming evidence [19]*19of defendant’s involvement in the crimes of which he was convicted”). Apart from defendant’s own statements, the state’s strongest evidence was M’s interview at CARES, which was given over a year after the alleged incident and after M had spoken with her mother about the incident. Defendant offered expert testimony about false memory and argued that the incident that the victim described had been planted by her mother. But questioning the CARES interview addressed only part of the state’s evidence.
The state’s closing argument repeatedly emphasized defendant’s own statements as proof that he was guilty. Thus, the determination of defendant’s guilt primarily depended on whether the jury believed defendant’s theory that his adjustment disorder, with its effect on his thought process and coping abilities, explains the incriminating statements. Callum’s testimony would have helped the jury assess defendant’s theory and “was evidence that defendant was entitled to present to the jury.” See Gherasim, 329 Or at 198; see also State v. Marquez-Vela, 266 Or App 738, 746, 338 P3d 813 (2014) (evidentiary error related to central factual issue in the case is less likely to be harmless). We do not suggest that the jury would have believed Callum’s testimony, but we cannot weigh that testimony, or assess the credibility of the other witnesses, in making the harmless error determination. See Davis, 336 Or at 32 (explaining that courts do not sit as factfinder when conducting harmless error review). Accordingly, we reverse and remand defendant’s conviction.7
Reversed and remanded.