Two of defendant’s great-grandchildren, five-year-old K and four-year-old A, separately reported that defendant had sexually abused them. Following a consolidated jury trial, defendant was found guilty on all charges related to A and not guilty on the single charge related to K. Defendant appeals the resulting judgment of conviction for two counts each of sodomy in the first degree, ORS 163.405, and sexual abuse in the first degree, ORS 163.427. First, he argues that the trial court erred by admitting A’s hearsay statements under OEC 803 (18a) (b) because the state failed to satisfy the notice requirements of that statute; we reject that assignment of error without discussion. Second, he assigns error to the trial court’s failure to strike, sua sponte, certain testimony by one of the state’s witnesses. The state elicited the testimony without defendant’s objection, but, according to defendant, it amounted to an impermissible comment on A’s credibility. For the reasons that follow, we conclude that defendant has not established that the trial court committed plain error. We thus affirm.
We state the facts necessary to provide context for the testimony at issue on appeal. K told her mother that defendant “tickles me down there [,]” pointing to her vagina. K’s mother took her to Liberty House, a child abuse assessment center, where she was interviewed. No charges were filed against defendant at that time. Subsequently, A told her mother that defendant made her put his penis in her mouth and touch his genitals. A’s mother filed a report with the police, who referred A to Liberty House. There, A repeated the abuse allegations to a Liberty House interviewer, Mills. A’s physical exam did not reveal physical signs of abuse. A police investigation ensued. The detective assigned to the case did not interview A because “[t]he interview had already been conducted by Liberty House [.]” Defendant was ultimately arrested and charged with various counts of sexual abuse and sodomy.
At trial, the victims and Mills testified on behalf of the state. On appeal, defendant challenges for the first time the trial court’s failure to strike a portion of Mills’s testimony. In evaluating whether the trial court plainly erred, *573“we must consider the entire context” in which the challenged comments were given. State v. Corkill, 262 Or App 543, 544, 325 P3d 796, rev den, 355 Or 751 (2014). We thus describe the trial proceedings involving Mills’s testimony in some detail.
Mills, a “nationally certified counselor,” stated that at Liberty House his “specialty or * * * expertise” was interviewing complainants. Mills estimated that he had interviewed “a couple hundred” children aged five or six years old. On direct examination, the prosecutor asked Mills to describe the specificity of A’s disclosure of abuse during his interview of her:
“[PROSECUTOR:] Was this a descriptive disclosure, not so descriptive disclosure? How would you describe it?
“A. I would — the—I’m in court and I’m being honest. The word that comes to my mind is this is a solid description from a child this age. I would say that her development seemed solid enough. And I have talked with kids of this age who have given less description than she did.”
On cross-examination, defendant questioned Mills, in part, concerning false accusations by children:
“[DEFENSE COUNSEL:] Okay. Now, you had testified that you have interviewed hundreds of children. And have you, through your experience of these interviews of hundreds of children, encountered incidences where the children had made false accusations against the person of interest?
“A. Because I’m in court and I just want to be clear, I have probably interviewed a total of 3(00) to 400 kids. Of that I can’t give you a definite number of, you know, how many are five or six. But I would say at least 100 to probably 200 just so I’m clear.
“In terms of kids making a false or — by false accusation; at Liberty House we don’t make judgments or determinations about statements. Whether in the end whoever that— whoever the end group is or person is outside of Liberty House that makes the determination if it’s true or not; we don’t do that. We simply gather whatever information we can, and then report back, like put it in the report and do these sorts of things.
*574“Q. And if you feel that misconduct has occurred, is this something that you would then refer to the District Attorney’s Office?
“A. We send our reports back to the referring agencies. And I think technically those agencies then refer on to the DA’s office.”
Defendant also attempted to question Mills about whether a child’s statement that he or she wanted to return to the alleged abuser is a factor in determining whether abuse has occurred:
“[DEFENSE COUNSEL:] What about a child, in your experience, have you had a child that says that they actually want to return to the people or the place where the alleged incidences occurred?
“A. Yes.
“Q. Okay. And is that a factor in your analysis of determining whether or not abuse did or did not occur?
However, the prosecutor objected:
“[PROSECUTOR:] Your Honor, I don’t think that he can talk about his analysis of whether or not something did or did not occur. I don’t think a witness can talk about whether or not they think something did or did not occur under [State v. Southard, 347 Or 127, 218 P3d 104 (2009)]. And that would probably open some doors to evidence defense counsel doesn’t want me to put in.”
Defendant then abandoned that line of questioning.
On redirect, Mills gave the testimony that defendant argues impermissibly commented on A’s credibility:
“[PROSECUTOR:] And as far as false accusations, based on your training and experience, statistically what is the average of false accusations?
“A. The exact numbers, from my experience, are unclear. Excuse me. What is clear is that children rarely lie about these things. And over — I believe it’s over 95 percent of the time when child — when a child discloses information about being sexually abused, it’s true.
“Q. Over 95 percent.
*575“A. I believe the statistic at Liberty House is 96 to 98 percent of the time when a child makes a disclosure about abuse, there’s truth to it. Some of those details that we talked about, maybe numbers of times or exactly where this happened when might be a little bit confusing for a child of [A’s] age. But—
“Q. The generality of abuse is—
“A. Generality, correct.”
In the state’s rebuttal to defendant’s closing argument, the prosecutor called the jury’s attention to the disputed testimony. In particular, the prosecutor argued that, “statistically,” there was a 95 percent chance that the victims were telling the truth. As noted above, the jury found defendant guilty on all counts related to A and not guilty on the single count related to K.
On appeal, defendant argues that Mills impermissibly vouched for A’s credibility when he testified that “the statistic at Liberty House is 96 to 98 percent of the time when a child makes a disclosure about abuse, there’s truth to it.” According to defendant, the trial court’s failure to strike that testimony was “an error of law apparent on the record” — i.e., plain error — that we should exercise our discretion to correct. See ORAP 5.45(1) (“No matter claimed as error will be considered on appeal unless the claim of error was preserved in the lower court * * *, provided that the appellate court may consider an error of law apparent on the record.”). The state counters that there is a “reasonable dispute about whether Mills vouched for A’s credibility” given his testimony that “at Liberty House we don’t make judgments or determinations about statements. Whether * * * it’s true or not; we don’t do that.”
An error is plain if it is a legal error that is obvious (i.e., not reasonably in dispute) and the court need not go outside the record or select among competing inferences to discern it. State v. Brown, 310 Or 347, 355, 800 P2d 259 (1990). Only if we conclude that an asserted error is plain do we determine whether to exercise our discretion to address the error. Ailes v. Portland Meadows, Inc., 312 Or 376, 382, 823 P2d 956 (1991). In this case, we agree with the state that the asserted error was not plain.
*576Our recent decision in Corkill is instructive. There, the prosecutor asked the defendant at various points in his testimony whether the police officer witnesses were lying; without objection, the defendant gave affirmative answers. Corkill, 262 Or App at 548-49. On appeal, the defendant argued that the trial court plainly erred in allowing vouching testimony, despite the lack of any defense objection. Id. at 549. We concluded that the trial court did not have a duty “to sua sponte preclude the prosecutor from asking whether prosecution witnesses were lying.” Id. at 544. We held that the defendant’s “unpreserved argument present [ed] no basis for reversal.” Id.
We began our analysis in Corkill by noting that,
“in most circumstances [,] it is beyond dispute that such questioning is objectionable, and that any responsive testimony commenting on the credibility of another witness would be inadmissible if it were objected to. That does not necessarily mean, however, that it is beyond dispute that a trial court must always intervene sua sponte to prevent that kind of evidence from reaching the jury.”
Id. at 552. In other words, we explained,
“[w]hen, as here, the claimed ‘plain error’ is associated with a trial court not having sua sponte interrupted a line of questioning (or not having excluded the resulting evidence sua sponte), the existence of any error does not depend solely on whether — as an abstract matter — the lawyer’s questions or the elicited answers would have been inadmissible if they had been objected to. Rather, any ‘plain error’ must relate to the trial court having not taken affirmative steps to intervene in the parties’ litigation.”
Id. at 551. We then offered the following summary of relevant precedent that is equally pertinent here:
“[T]he cases in which we have held that trial courts are obligated to exclude evidence in the absence of an objection — and plainly err if they do not — all share an important characteristic that is not present here.
“Each of the cases in which we have held that a trial court should have excluded evidence sua sponte *** has involved true ‘vouching’ evidence, that is, one witness’s testimony that he or she believes that another witness is or is *577not credible, which a party offers to bolster or undermine the veracity of that other witness.”
Id. at 552.
In this case, Mills’s testimony did not run afoul of the prohibition on true vouching when he testified, “I believe the statistic at Liberty House is 96 to 98 percent of the time when a child makes a disclosure about abuse, there’s truth to it.” Mills stopped short of stating that A was like the 96 to 98 percent of Liberty House complainants whose reports were truthful. Cf. State v. Remme, 173 Or App 546, 558, 23 P3d 374 (2001) (stating that it is impermissible for a witness to “‘connect the dots’ explicitly by [giving] testimony relating [a general] phenomenon’s specific application to a particular witness’s testimony”).
Although “statements that fall short of such overt vouching also may be impermissible,” State v. Lupoli, 348 Or 346, 357, 234 P3d 117 (2010), it is not apparent that Mills even indirectly connected the statistic that he cited to A’s report of abuse or its purported veracity. Instead, the testimony by Mills more closely resembles the kind of “general information *** indicating] that [a general class of witnesses] is or is not” credible that we approved in State v. Preuitt, 255 Or App 215, 221, 296 P3d 648, rev den, 353 Or 868 (2013).
In short, “because the evidence in question here was not clearly vouching for [A’s] credibility * * * but was more akin to permissible testimony,” State v. Wilson, 266 Or App 481, 495, 337 P3d 990 (2014), we agree with the state that there is a reasonable dispute as to whether Mills impermissibly vouched for A. Thus, the trial court did not plainly err by not sua sponte striking the disputed testimony at issue here.
Affirmed.