Defendant appeals a judgment of conviction for six counts of sexual abuse in the first degree, ORS 163.427, and six counts of sodomy in the first degree, ORS 163.405. In a prior appeal, we vacated defendant’s earlier judgment of conviction in light of State v. Southard, 347 Or 127, 218 P3d 104 (2009), and remanded the case to the trial court. Defendant was then retried, convicted on fewer counts than in the first trial, and sentenced to 844 months in prison—substantially longer than his original sentence of 450 months. On appeal, defendant assigns error to (1) the trial court’s denial of his motion to dismiss on double-jeopardy grounds, (2) certain evidentiary rulings at trial, and (3) the imposition of a harsher total sentence in the absence of express findings by the trial court to justify the disparity.
As explained below, we reject defendant’s double-jeopardy and evidentiary challenges. We conclude, however, that the trial court plainly erred by imposing a sentence in excess of defendant’s original sentence without making findings on the record to permit a conclusion that the sentence was not the product of vindictiveness. We therefore remand for resentencing.1
FACTS
We limit our discussion to those facts that are relevant to this appeal. The counts of conviction pertain to three victims—M, C, and L. Defendant’s brother and his brother’s wife are the adoptive parents of M, C, and L, along with several other children who joined the family by adoption or by birth. M, C, and L are not birth siblings and came into the Criswell home at different times. Starting in October 2004, the three children were regular visitors to the home of defendant and his wife, sometimes spending Saturday nights there. In December 2006, M, C, and L were spending time in M’s bedroom outside the presence of their parents. M yelled for her mother to come into the room, and C told her mother that defendant had sexually abused her. Mother *149called father into the room, and C repeated the statement. M told her mother that she had previously tried to tell her parents that defendant was doing inappropriate things to C. M told her parents that she herself had “always fought [defendant] off.” The parents asked L whether anything had happened to her, and L did not say anything initially. A few days later, L communicated to her mother that defendant had sexually abused her as well.2
Mother contacted the Department of Human Services (DHS) two days after C’s disclosure. Several days later, a detective and a DHS worker interviewed some of the Criswell children at school. During the interview with C, she told them that defendant had abused her approximately “50 times,” and that she stated that she felt something “wet” during the abuse. Neither the detective nor the DHS worker could recall who introduced the term “wet.”
During the trial, M testified that defendant touched her thighs, legs, chest, stomach, arms, back, neck, and kissed the top of her head. M stated that the touching was over and under her clothing. M also testified that she saw defendant touching C on several occasions. C testified that defendant touched her inappropriately on her thighs and stomach. C testified that defendant would lie behind her and touch her inappropriately, and she felt something go inside her “butt.” C stated that she did not think defendant touched her chest, and that she could not remember whether there was anything “wet.” L testified that defendant touched her chest under her clothes and that defendant placed her hand on his “private area.”
At trial, defendant testified on his own behalf. He denied engaging in any inappropriate contact with M, C, or L.
The jury convicted defendant of one count of sexual abuse in the first degree with respect to M, six counts of sodomy in the first degree with respect to C, and five counts of sexual abuse in the first degree with respect to L. The trial court sentenced defendant to 844 months’ imprisonment.
*150DOUBLE JEOPARDY
In his first assignment of error, defendant argues that the trial court erred in denying his motion to dismiss on double-jeopardy grounds. Defendant argues that retrial was barred because official misconduct caused the preceding mistrial.
We begin with additional facts relevant to defendant’s motion to dismiss. In the weeks following the disclosures, M, C, and L were evaluated at Juliette’s House, a child abuse intervention center. Dr. Moore, a medical doctor, performed physical examinations on both C and L. During C’s examination, C described various forms of abuse by defendant, including penetration of her anus. C also stated that, during the abuse, she “felt something wet.” Moore’s examination of C revealed physical findings consistent with abuse. Moore also examined L, and her physical exam was normal. Dr. Miller, the medical director of Juliette’s House, performed M’s examination. M allowed only a limited physical exam. Warner, a forensic interviewer, conducted interviews of M, C, and L as part of the assessment process at Juliette’s House.
Defendant’s first trial occurred in 2008. The jury convicted defendant on 54 of 64 counts. We vacated and remanded the case for reconsideration in light of Southard, 347 Or at 142 (holding that a physician’s diagnosis of sexual abuse in the absence of physical evidence is inadmissible under OEC 403).3
Defendant’s second trial ended in a mistrial for reasons not pertinent on appeal. The third trial started five days later. As the prosecutor prepared to question Warner, the interviewer from Juliette’s House, the prosecutor alerted the court that she did not have a final, redacted version of the Juliette’s House reports. The prosecutor indicated that she “may still hand a copy of [the reports] to Ms. Warner to just identify that it’s her signature and it’s her copy of *151the report.” Defense counsel stated that he had “no problem with that,” so long as “Ms. Warner has been well-apprised of what to say and *** what subjects to stay away from[.]” The prosecutor replied, “I think [Warner] understands that we’re not to talk about the diagnosis.”
The parties and the court then discussed whether Warner and Moore would be allowed to testify that counseling had been recommended for M, C, and L. Defendant argued that such testimony should not be allowed because it created a “very strong inference” that the children had been diagnosed with sexual abuse. The trial court agreed and ruled that testimony about treatment recommendations for M and L would not be permitted, leaving the question open with respect to C.
On the sixth day of trial, the state called Miller. Miller described M’s assessment process at Juliette’s House:
“It had specifically entailed history gathering, finding what the concerns were regarding [M], taking a detailed past medical history, social history, doing a complete physical examination. Part of our assessment involves an interview that I observe but am not—play an active role in. And using that information, then, to formulate a diagnosis and treatment plan.”
Defendant did not object to Miller’s reference to “a diagnosis and treatment plan.” Shortly afterward, the following exchange occurred:
“[PROSECUTOR]: Doctor, given the *** type of touching or abuse that [M] was describing, would you have expected to have any abnormal findings upon a physical exam?
“[MILLER]: No.
“ [PROSECUTOR]: Why not?
“ [MILLER]: What she was disclosing was really touching or fondling. And it would be very unlikely for that to leave any lasting signs.
“[PROSECUTOR]: I’m going to show you what has been marked as State’s Exhibit 15, and ask you if you can identify this item.
*152“[MILLER]: That’s a copy of my report from that assessment date in December of 2006.
“[PROSECUTOR]: Does this report contain one part, or more than one part?
“[MILLER]: Contains two parts. It contains the part that I dictated that has the medical history and physical examination. It has a part that Michele Warner dictated, which is a social history and the interview summary. And then I take those two pieces of information and formulate a diagnosis and treatment plan.”
(Emphasis added.) Defendant immediately moved for a mistrial based on Miller’s reference to “formulating] a diagnosis and treatment plan.” The court denied defendant’s motion and instructed the jury to disregard Miller’s testimony “referring to the content of her report.” The following day, defendant renewed his motion for a mistrial. After the court received questions from several jurors expressing confusion about which portion of Miller’s testimony they should disregard, the court granted defendant’s motion.
Before the fourth trial began, defendant filed a motion to dismiss on double-jeopardy grounds, citing the exchange quoted above between the prosecutor and Miller in the third trial. Defendant’s theory was that both the prosecutor and Miller had acted with indifference to the risk of a mistrial, amounting to misconduct that prohibited the state from trying defendant again. Defendant reasoned that the prosecutor must have informed Miller “not to say anything about diagnosis, diagnostic findings, [and] treatment recommendations” because “that’s why this case was reversed in the first place.” Defendant argued:
“[T]o say that somehow, a trained, experienced agent of the State, coming back to testify * * * on the exact same thin ice—that brought this case back in the first place and * * * 40 minutes before, 20 minutes before, [we] were very clear with all three of us that this couldn’t be talked about. Took a break so that the * * * witness could, again, be instructed. * * * For her to then blurt that out I think clearly could be indifferent to the consequence.”
Defendant further argued that it was misconduct for the prosecutor to hand Miller an unredacted copy of the Juliette’s *153House report, which had headings including “DIAGNOSTIC FINDINGS” and “TREATMENT RECOMMENDATION.”
The prosecutor responded that the purpose of showing the report to Miller had been simply to identify it and not to discuss its contents. The prosecutor added that Miller was “very well aware” of the reasons for reversal after the original trial, and she knew that “she would not be testifying to her diagnosis in this case.” The prosecutor argued that Miller’s response to the question was “automatic” and that Miller was “talking in general terms, not saying what the diagnosis was in this case.”
The trial court denied defendant’s motion to dismiss. The court reasoned that the law “at the time” would not have necessarily required a mistrial based on Miller’s testimony. The court found that, “at best,” it was “negligent” for Miller to make the disputed statement, reasoning that “she’s a doctor testifying the way doctors do,” that “terms like ‘diagnosis’ and ‘treatment plan’ are everyday medical language,” and that Miller did not “realize [] what she was doing.” With respect to the prosecutor, the court found that the prosecutor did not intend to cause a mistrial, nor was she indifferent to the risk of causing a mistrial. The trial court went on:
“In fact, I would say rather the opposite, that all of us, * * * did not want to have the case retried. And I would say that for [defendant], as well as the victims in this case, and not to mention the attorneys, but nobody wanted a mistrial in this case. So I don’t think there’s any indication that the State was indifferent.”
We review a trial court’s ruling on a motion to dismiss on double-jeopardy grounds for legal error. State v. Garner, 234 Or App 486, 491, 228 P3d 710, rev den, 348 Or 621 (2010). We are bound by the trial court’s findings of historical fact if there is evidence in the record to support those findings. State v. Mays, 269 Or App 599, 601, 346 P3d 535, rev den, 358 Or 146 (2015) (citing State v. Ehly, 317 Or 66, 74-75, 854 P2d 421 (1993)).
Article I, section 12, of the Oregon Constitution provides that “[n]o person shall be put in jeopardy twice *154for the same offence [sic].” When a defendant moves for a mistrial or otherwise consents to termination of the trial, he is generally considered to have waived double-jeopardy protections, unless his decision to move for a mistrial is the product of certain forms of official misconduct. Mays, 269 Or App at 616. In such cases, retrial is barred based on the “‘theory that the official misconduct put the defendant to the “Hobson’s choice” of either moving for a mistrial or continuing with a jury which was, arguably, favorably disposed to the defendant but is now contaminated by the misconduct.’” Id. (quoting State v. Kimsey, 182 Or App 193, 203, 47 P3d 916 (2002)). The objective of Oregon’s double-jeopardy protections is to “protect defendants against the harassment, embarrassment [,] and risk of successive prosecutions for the same offense,” and it is not “a sanction to be applied for the punishment of prosecutorial or judicial error.” State v. Kennedy, 295 Or 260, 272-73, 666 P2d 1316 (1983) (internal quotation marks omitted).
Consequently, Article I, section 12, bars a retrial when the three following conditions are met: (1) the “improper official conduct is so prejudicial to the defendant that it cannot be cured by means short of a mistrial”; (2) “the official knows that the conduct is improper and prejudicial”; and (3) the official “either intends or is indifferent to the resulting mistrial or reversal.” State v. Pratt, 316 Or 561, 579, 853 P2d 827 (1993) (internal quotation marks omitted). The standard requires “‘some conscious choice of prejudicial action before the [double-jeopardy] guarantee bars correction of the error by a new trial. Negligent error, gross or otherwise, is not enough.’” Kimsey, 182 Or App at 203-04 (brackets in original; internal quotation marks omitted). In determining whether Article I, section 12, precludes a retrial, “the dispositive issue is whether the [official] had the requisite scienter” See Mays, 269 Or App at 617.
Here, we need not determine whether the first two prongs of the inquiry are satisfied, because the third prong is dispositive. The trial court expressly found that neither Miller nor the prosecutor was indifferent to the risk of a mistrial. Accordingly, we are bound by those findings as long as there is evidence to support them. Id. at 618. We *155ordinarily defer to a trial court’s credibility findings when “credibility based on demeanor is a crucial factor” in making a factual determination. Id. (internal quotation marks omitted). Here, the trial court was in the best position to assess the demeanor of both Miller and the prosecutor. And, as explained further below, there is evidence to support the trial court’s findings that neither of them was indifferent to the risk of a mistrial.
With respect to the prosecutor, defendant relies heavily on the fact that she did not redact the Juliette’s House report before handing it to Miller. However, earlier in the trial, defendant had expressly agreed that the prosecutor could question Warner about an unredacted version of a Juliette’s House report for identification purposes so long as Warner was aware of what topics were impermissible. That evidence supports a finding that the prosecutor would have understood the same conduct to be acceptable with Miller. In addition, the prosecutor stated that she expressly instructed Miller not to discuss the diagnosis. Thus, there is evidence to support the trial court’s finding that the prosecutor was not indifferent to the risk of a mistrial.
Similarly, evidence supports the trial court’s finding that Miller was not indifferent to the risk of a mistrial.4 The trial court’s finding with respect to Miller rested in large part on its conclusion that her testimony was not a “direct” violation of the law. The trial court correctly noted that, at the time of Miller’s testimony, it was not clear that a reference to the mere existence of a diagnosis and treatment plan necessarily crossed the line into improper vouching testimony. Cf. State v. Williams, 249 Or App 514, 515, 278 P3d 69 (2012) (holding that it was not plain error for the trial court to allow a witness to “mention [] that she gave the child’s father recommendations for the child’s future health and safety”). Thus, based on the state of the law at the relevant time, the trial court correctly concluded that it was not obvious that a motion for a mistrial had to be granted based *156on Miller’s testimony. That supports the trial court’s finding that Miller was not “indifferent” to that risk.
In sum, we conclude that the trial court did not err in denying defendant’s motion to dismiss on double-jeopardy grounds. Because we defer to the trial court’s findings that neither Miller nor the prosecutor was indifferent to the risk of a mistrial, we reject defendant’s assignment of error under Article I, section 12. Accordingly, we also reject defendant’s argument under the Fifth and Fourteenth Amendments to the United States Constitution because, under federal law, retrial is barred only when an official engages in conduct intended to provoke a mistrial. Oregon v. Kennedy, 456 US 667, 675-76, 102 S Ct 2083, 72 L Ed 2d 416 (1982); see also Kennedy, 295 Or at 271-76 (discussing and distinguishing the two standards).
EVIDENTIARY CHALLENGES
Defendant’s second through ninth assignments of error challenge several evidentiary rulings by the trial court. The challenged rulings fall into two categories. First, defendant objected to the trial court’s admission of seven statements that, according to defendant, amounted to impermissible “vouching.” Second, defendant argues that the court erred in refusing to admit evidence that C was found to have crab lice in her eyebrows when she came to the Criswell home at the age of four (six years before she accused defendant of abuse). We reject the latter argument without discussion and discuss below defendant’s objections to what he characterizes as “vouching” evidence.
It is well established that “one witness may not comment on the credibility of another witness.” State v. Chandler, 360 Or 323, 330, 380 P3d 932 (2016). That prohibition applies to “direct comments on the credibility of another witness, as well as to statements that are ‘tantamount’ to stating that another witness is credible.” Id. at 331 (quoting State v. Beauvais, 357 Or 524, 543, 354 P3d 680 (2015)). The “rule applies to credibility opinions about statements that a witness made either at trial or on some other occasion.” Id. We review for legal error whether the trial court admitted impermissible vouching evidence. State v. Marquez-Vela, 266 Or App 738, 739, 338 P3d 813 (2014).
*157We first address certain testimony provided by Warner, the Juliette’s House interviewer. During mother’s testimony, the following exchange took place:
“[PROSECUTOR]: Do you recall [M] coming to you at any time before December of 2006, and telling you about [defendant] doing inappropriate things with her?
“ [MOTHER]: No, not at all.
“[PROSECUTOR]: In hindsight, is there anything that [M] ever said to you during those years, 2001 up until December 2006, that could have been her trying to tell you?
“[MOTHER]: There was a—I can remember a time when she would come up and say, ‘[Defendant] was extra weird this weekend,’ something along the lines, ‘ [Defendant] was really weird this weekend.’ * * * [S]he didn’t clarify, and I didn’t go any further, because it seemed like it was not a big deal, because she just said, ‘[Defendant] was, you know, extra weird, too weird this weekend,’ or something.”
(Emphases added.)
Testifying next, Warner said that she believed that she had “specialized knowledge about the actions and reactions of victims that would help [the] jury in understanding the actions of a child who’s been a victim of sex abuse.” The following exchange took place:
“[PROSECUTOR]: Based on your training and experience, do children sometimes attempt to tell in a way that adults don’t understand what they’re saying?
“[WARNER]: Yes. Frequently in interviewing young children, they’ll say, T—I tried to tell, or I told my mom, and she didn’t believe me, or she didn’t listen, or whatever.’ And sometimes if I ask the child, ‘Well, what did you say to your mom?’ They may say something like, 'Well, I told her that he’s weird’ or that, T don’t like going to his house,’ or just something * * * that’s pretty general like that.”
(Emphasis added.)
Defendant objected to the testimony, arguing that Warner, by describing what a child “may say” in nearly identical terms to those that mother used to describe what M actually did say, was vouching for M’s credibility. The trial *158court recognized that Warner’s statement was “either identical or close to identical” to how mother described M’s statements, but found that Warner’s testimony was consistent with “her expertise as an interviewer.” Accordingly, the court overruled defendant’s objection. On appeal, defendant argues that Warner’s testimony constituted impermissible vouching because “Warner told the jury that this is how children who have really been abused behave, and M behaved that way.”
“Expert testimony that provides jurors with useful information in making their own credibility assessment ordinarily is admissible, as long as it is not either a direct comment on the credibility of a witness or tantamount to a direct comment on the credibility of a witness.” Beauvais, 357 Or at 545. Accordingly, an expert witness may describe characteristics typical of an abuse victim and then apply that understanding to the case at hand, so long as that testimony “d[oes] not rely on an assessment of the victim’s credibility.” See State v. Swinney, 269 Or App 548, 559, 345 P3d 509, rev den, 357 Or 743 (2015) (upholding admission of testimony “objectively describing characteristics typical of grooming in family sex abuse cases,” and then applying that knowledge to the behaviors described by the victim); see also State v. Sundberg, 268 Or App 577, 583-85, 342 P3d 1090, rev den, 357 Or 325 (2015) (holding that trial court should have admitted a doctor’s statements that the complainant’s description of her anatomy was typical of children her age, thereby providing the jury with an explanation as to why the complainant may have described the abuse in inconsistent ways). The inquiry turns upon whether the expert testimony assists, rather than usurps, the jury’s own assessment of the complainant’s credibility. See Beauvais, 357 Or at 545. Important to this inquiry is whether the testimony can assist the jury in interpreting behavior or circumstances that are “beyond the ordinary experience of a lay finder of fact.” Id.; see also State v. Middleton, 294 Or 427, 435-36, 657 P2d 1215 (1983) (upholding the admission of testimony about victim recantation because “[explaining this superficially bizarre behavior by identifying its emotional antecedents could help the jury better assess the witness’s credibility”).
Contrary to defendant’s argument on appeal, the trial court did not err in admitting Warner’s statements *159because her testimony “‘d[id] no more than provide jurors with useful, nonconclusive information from which inferences as to credibility may be drawn.’” Sundberg, 268 Or App at 584 (quoting State v. Remme, 173 Or App 546, 562, 23 P3d 374 (2001) (emphases in Remme)). It is true that Warner’s testimony may have provided the jury with information from which it could infer that M’s descriptions of abuse were credible. That is, the jury might have combined (1) Warner’s testimony that a child might say “he’s weird” as a way to describe abuse with (2) mother’s testimony that M said precisely that about defendant to (3) infer that M made an attempt to disclose her abuse to her mother. From that, the jury might have inferred further that, if M made an attempt to disclose abuse, then the abuse likely occurred. However, Warner’s testimony did not cross the line into impermissible vouching because Warner commented, at most, on whether M’s statements may have been an attempt at disclosure, not on whether that disclosure was credible. Cf. State v. Keller, 315 Or 273, 278, 285, 844 P2d 195 (1993) (holding that it was error for the trial court to admit a doctor’s testimony that “[t]here was no evidence of leading or coaching or fantasizing” in a complainant’s interview (brackets in original)); State v. Milbradt, 305 Or 621, 626, 629-30, 756 P2d 620 (1988) (holding that it was error for the trial court to admit a psychotherapist’s testimony that, among other things, complainant did not display “indicators of deception”). Thus, Warner’s testimony gave the jury useful information as to how it should evaluate M’s purported statements and did not supplant the jury’s role in assessing M’s credibility. Accordingly, the trial court did not err in allowing Warner’s testimony.
We next turn to defendant’s arguments that the trial court erred in admitting certain statements in the Juliette’s House reports. On appeal, as before the trial court, defendant argues that referring to M as “present [ing] as a competent fourteen-year-old” and C as “presenting] as a competent *⅝* ten-year-old,” amounts to vouching for the ability of M and C to provide “accurate, reliable information.” Defendant further argues that describing L as presenting as “a reluctant and anxious, but competent seven-year-old” implies that L “provided information only because she was *160required to do so, making it more likely to be truthful.” We conclude that those descriptions do not cross the line into impermissible vouching because they are demeanor-based observations that provide the jury with useful information from which it could independently assess the complainants’ credibility. See State v. Wilson, 266 Or App 481, 490, 337 P3d 990 (2014), rev den, 356 Or 837 (2015) (“[T]estimony about the physical appearance of a speaker, or testimony that is solely descriptive of the manner in which a communication is made—so-called demeanor evidence—is admissible and is not vouching evidence.”); see also State v. Lupoli, 348 Or 346, 362, 234 P3d 117 (2010) (reasoning that expert testimony stating a complainant’s statements were “developmentally appropriate for her age is the kind of expert opinion that can assist a jury and ordinarily would be admissible,” while excluding such evidence because it was “inextricably bound up with” testimony that constituted “clear vouching” (internal quotation marks omitted)).5
Defendant also argues that the trial court erred in refusing to strike mother’s statement explaining that she and her husband waited two days to contact DHS after C’s disclosure because “it just took us that long to really sink in that this was really happening.” The trial court did not err. When viewed in context, the disputed statement was made for the relevant and proper purpose of explaining mother’s response to the complainants’ disclosures; in that light, the relevance of the statement did not depend on whether mother believed that the disclosures were true. See State v. Hunt, 271 Or App 347, 350, 350 P3d 521 (2015) (concluding that the trial court did not plainly err in failing to strike a deputy’s testimony that he did not originally believe the victim because it was admitted to show the deputy’s state of mind regarding the investigation). Cf. Chandler, 360 Or at 334 (“[A] court does not err in admitting an out-of-court statement as to the credibility of a witness or nonwitness complainant if the statement is offered for a relevant, *161non-opinion purpose”). Thus, we conclude that the trial court did not err in its evidentiary rulings because none of the statements challenged by defendant constituted impermissible vouching.
SENTENCING
Finally, we turn to defendant’s argument that his sentence must be vacated. After defendant’s first trial, the jury convicted him on 54 counts, and the trial court imposed a sentence of 450 months. After defendant’s fourth trial, the jury convicted him on 12 counts, and the trial court— with a different judge presiding—imposed a sentence of 844 months. That judge did not make findings on the record to explain the discrepancy between the two sentences, and defendant did not object to the sentence or request that the court make such findings. On appeal, defendant argues that the sentence is presumptively vindictive and that we should review and correct that error as “plain error.”
A trial court is permitted to “impose a harsher sentence on a defendant after retrial or remand, as long as the court’s grounds for doing so are not directed at punishing the defendant for appealing his or her original convictions or sentences.” State v. Partain, 349 Or 10, 26, 239 P3d 232 (2010); see also State v. Robledo, 282 Or App 96, 101, 386 P3d 136 (2016) (“[V]indictiveness against a defendant for having successfully attacked his first conviction must play no part in the sentence he receives after a new trial[.]” (Internal quotation marks omitted.)). In Partain, the Supreme Court adopted the following standard for cases subject to resentencing:
“If an Oregon trial judge believes that an offender whom the judge is about to resentence should receive a more severe sentence than the one originally imposed, the judge’s reasons must affirmatively appear on the record. Those reasons must be based on identified facts of which the first sentencing judge was unaware, and must be such as to satisfy a reviewing court that the length of the sentence imposed is not a product of vindictiveness toward the offender. Absent such facts and reasons, an unexplained or inadequately explained increased sentence will be presumed to be based on vindictive motives, and will be reversed.”
*162Id. at 25-26; see also State v. Febuary, 274 Or App 820, 828-29, 361 P3d 661 (2015) (stating that “the standard for measuring whether a new sentence triggers the presumption of vindictiveness is the aggregate approach”—meaning that the relevant inquiry is whether the trial court sentenced defendant to a “more severe total sentence”). Based on Partain, defendant argues that he is entitled to resentencing because the trial court offered no explanation for the sentence, nor does the record reveal additional facts that would support the imposition of a harsher sentence.
We first conclude that the trial court’s error is “plain.” See State v. Inman, 275 Or App 920, 927, 366 P3d 721 (2015), rev den, 359 Or 525 (2016) (“First, we determine whether the trial court plainly erred”). An error is “plain” if “(1) the error is one of law, (2) the error is obvious, not reasonably in dispute, (3) the error appears on the face of the record, * * * and the facts constituting the error are irrefutable.” Id. (internal quotation marks omitted). To begin, whether the trial court satisfied Partairís prerequisites is one of law. State v. Sierra, 278 Or App 96, 100, 374 P3d 952 (2016). Second, the error is obvious. Partain’s requirements are clear, and the trial court did not identify any “facts of which the first sentencing judge was unaware” to explain its reasons for imposing a sentence almost double the length of the original sentence. Third, the trial court’s failure to make the required findings is apparent from the sentencing transcript. Neither the state nor the sentencing judge pointed to any new facts that were presented during the fourth trial and not the first trial. The extent of the trial court’s explanation for defendant’s sentence was that the state’s arguments had “logic” behind them, that a lengthy sentence would provide “accountability for time periods” and for “harm against specific victims,” and those considerations were “intertwined with the application of mandatory minimum sentences.” Thus, the record divulges nothing that would justify the considerably increased sentence, and the state does not argue otherwise.
Even though we conclude that the error is plain, we must also consider “whether we should exercise our discretion to correct that error.” Inman, 275 Or App at 927. In so doing, we consider “an array of considerations, such as the *163competing interests of the parties, the nature of the case, the gravity of the error, and the ends of justice in the particular case.” Id. (internal quotation marks omitted). When reviewing a sentencing error, we particularly consider “whether a defendant encouraged the trial court’s imposition of the erroneous sentences, the possibility that the defendant made a strategic choice not to object to the sentences, the role of other sentences in the case, and the interests of the judicial system in avoiding unnecessary and repetitive sentencing proceedings.” State v. Welsh, 267 Or App 8, 14, 340 P3d 132 (2014) (internal quotation marks omitted).
The state argues that the trial court’s failure to make findings explaining defendant’s sentence is “unreviewable,” pointing to the Supreme Court’s decision in State v. Bucholz, 317 Or 309, 855 P2d 1100 (1993). In Bucholz, the defendant assigned error to the trial court’s failure to make findings to support its decision to impose consecutive sentences as provided in former ORS 137.123(4) (1991), amended by Or Laws 1995, ch 657, § 2, renumbered as ORS 137.123(5)(b) (1995). Id. at 319-20. The defendant was convicted of both theft and unlawful delivery of methamphetamine to a minor, id. at 311, and former ORS 137.123(4)(b) (1991) permitted a court to impose a consecutive sentence if the court found that an offense “caused or created a risk of causing greater or qualitatively different loss, injury or harm to the victim or caused or created a risk * * * to a different victim.”
The Bucholz court concluded that, even assuming the error otherwise satisfied the requirements for plain-error review, it was “not at all clear that the evidence in this case concerning the nature of the offenses sentenced consecutively would fall short of satisfying the requirements of [former] ORS 137.123(4)(b) [(1991)].” Id. at 321. The court continued:
“Defendant would have an appellate court reverse for absence of findings even though, had the matter been called to the sentencing court’s attention, applicability of paragraph (4)(b) might easily have been established. To preserve an error in the face of the possibility that the statute expressly permits consecutive sentencing in the situation at hand, a defendant who objects to [a] lack of express *164findings to bring the case under subsection (4), must place that objection on the record at the time of sentencing.”
Id. Thus, Bucholz relied substantially on the fact that the required findings might easily have been made. We later explained that Bucholz stands for the proposition that, even if the trial court plainly erred, in those circumstances, it was not appropriate for the court to exercise its discretion to correct any error. State ex rel Dept. of Human Services v. J. N., 225 Or App 139, 145, 200 P3d 615 (2009).
Here, unlike in Bucholz, we cannot conclude that findings necessary to satisfy Partain might have been easily established. The state does not point to any factual development in the fourth trial that could readily provide a non-vindictive explanation for the discrepancy between the two sentences. Cf. Welsh, 267 Or App at 14 (declining to follow Bucholz because “if the matter had been brought to the trial court’s attention, the trial court could not have made the necessary findings” (emphasis added)). Given that the record of the fourth trial lacks any obvious factual basis that would justify the increased sentence, there is no reason to believe that defendant made a strategic choice not to object to the sentence, nor does the record show that defendant encouraged the court’s failure to make findings to support the sentence. We also note that, if, indeed, the trial court imposed a longer sentence for vindictive reasons—which, under Partain, we presume it did absent an explanation to the contrary—the gravity of the error is severe, resulting in more than 30 additional years of imprisonment. For that same reason, we conclude that the ends of justice are served by remanding the matter to the trial court for resentencing. Accordingly, we exercise our discretion to correct the sentencing error.
Remanded for resentencing; otherwise affirmed.
Our disposition obviates the need to consider defendant’s contention that the court plainly erred in ordering him to pay $5,000 in court-appointed attorney fees.
At the time of the disclosures, M was 14 years old, C was 10 years old, and L was 7 years old.
At defendant’s first trial, Moore and Miller testified regarding their medical diagnoses of sexual abuse as to all three victims, even though only C had physical signs of abuse. The state conceded on appeal that that evidence was inadmissible under Southard and that a remand was required.
For purposes of this analysis, we assume without deciding that the testimony of Miller, a Juliette’s House doctor, may independently qualify as “official conduct” for purposes of the double-jeopardy analysis.
We also conclude that the trial court did not err in declining to strike the following statement from the Juliette’s House report pertaining to C: “She also gives a description of multiple episodes of ejaculation by [defendant].” This statement does not directly or implicitly vouch for the credibility of C’s account of the abuse. Instead, it constitutes the use of a technical term for what C described, if true.