¶ 1 Defendant, Israel Heredia-Cobos, appeals the judgment of conviction entered on a jury verdict finding him guilty of sexual assault on a child. One of the issues he raises is whether the district court abused its discretion in allowing a forensic interviewer to testify that the child victim, Y.P., didn't appear to have been coached as to what to say during an interview. Though such testimony is usually inadmissible, we conclude that in this case the defense opened the door to the forensic interviewer's testimony by challenging the victim's statements on the basis that she had fabricated them, at least in part, because of coaching by her relatives and others. We also reject defendant's other contentions, and therefore we affirm.
I. Background
¶ 2 Defendant is Y.P.'s great uncle. When Y.P. was nine years old, she spent the night at defendant's home with other family members. While she was playing on the trampoline *863with her cousins, defendant came outside and said that cake was being served. After Y.P.'s cousins went inside the house, defendant pushed Y.P. down, got on top of her, and touched her breast. He then tried to put his hand down her pants, but Y.P. pushed him off and ran inside.
¶ 3 Y.P. reported the assault just over four years later after a classmate told her that she had been raped by her father. The People charged defendant with a single count of sexual assault on a child.
II. Discussion
¶ 4 Defendant contends that the district court abused its discretion by (1) allowing the forensic interviewer who had interviewed Y.P. to testify that Y.P. didn't show any signs of having been coached and (2) allowing evidence of his prior acts of a sexual nature involving other relatives in violation of CRE 404(b).
A. Witness Testimony Regarding Coaching
1. Preservation and Standard of Review
¶ 5 The parties agree that defendant preserved this issue.
¶ 6 We review the district court's evidentiary rulings for an abuse of discretion. People v. Faussett , 2016 COA 94M, ¶ 33, 409 P.3d 477. To constitute an abuse of discretion, the district court's evidentiary ruling must have been manifestly arbitrary, unreasonable, or unfair, or based on a misunderstanding or misapplication of the law. Id.
2. Analysis
¶ 7 The prosecutor called Lisa Tani, a forensic interviewer, to testify as an expert about her interview of Y.P. (Defense counsel didn't object to the prosecutor's request that Ms. Tani be allowed to give expert testimony.)
¶ 8 Ms. Tani initially testified about how she interviews children generally. The prosecutor asked her whether "there are certain things that you are looking for or precautions that you are taking throughout the interview?" Ms. Tani replied, "So while we interview children we assess for coaching, suggestibility, ... how trauma may affect their memory, ... development, those type of things."
¶ 9 The prosecutor later returned to the subject of coaching.
Q: Okay. You also said that you're looking for signs of suggestibility, um, and you mentioned coaching, signs of coaching. What signs of coaching are you talking about?
A: So if a child has been coached, which typically you will see a child being coached under the age of 10, they are usually coached on that specific event. So a caregiver-they might overhear someone talking about an event, so they will come in and just tell you about that event. Typically, they don't have the sensory detail that we look for, the peripheral details, and they have limited information regarding that event.
¶ 10 Ms. Tani also testified that she doesn't assess the child's credibility, but when asked what she assesses during an interview she said, "I will assess on coaching. If I feel that ... they were being suggestive, ... I will come up and talk to the parents."
¶ 11 The prosecutor then turned to Ms. Tani's interview of Y.P.
Q: Did you-and I'm not asking you to opine on her credibility, but did you see any indication throughout that interview of the concerns that you have just talked about?
Defense counsel objected that the question called for Ms. Tani to comment on Y.P.'s credibility. The prosecutor argued that such testimony wasn't an opinion about credibility and pointed out that the defense had brought up "numerous times" at trial that Y.P. had made up the allegations "because she overheard gossip or ... was somehow trying to fit in as a peer at school." So, the prosecutor argued, the testimony was relevant to rebut that theory. The court overruled the objection, reasoning that "whether signs of coaching were there [doesn't go] directly to credibility or truthfulness."
¶ 12 The prosecutor then continued questioning Ms. Tani about whether Y.P. appeared to have been coached.
Q: Okay. Ms. Tani, as I said, um, you can't opine on the credibility of the *864child or whether you believed the child or things like that. What I'm asking you is based on the things that you just talked about that you're assessing during an interview, did you see any indications of coaching during this interview?
A: No, I did not.
Q: Or that things had been suggested?
A: No, I did not.
Q: You also talked about, um, looking for sensory details and peripheral details throughout the interview. Did you observe [Y.P.] make details that are peripheral in this case?
A: Yes, I did.
¶ 13 Ms. Tani then testified as to the various details Y.P. told her about during the interview, such as whom she was with, what she was doing (playing on the trampoline), the smell of defendant's breath, and how cold it was outside.
¶ 14 It doesn't appear that defendant challenges Ms. Tani's testimony about how she interviews children, including that she looks for signs of coaching. Instead, he focuses on Ms. Tani's testimony that she didn't see any indications that Y.P. had been coached. We conclude that although such testimony ordinarily is improper (because it's tantamount to vouching for the child's credibility), in this case the testimony was admissible to rebut defendant's theory of defense. See People v. Quintana , 882 P.2d 1366, 1375 (Colo. 1994) (erroneous admission of evidence under a rule of evidence not reversible where evidence was admissible for a different reason); People v. Pernell , 2014 COA 157, ¶ 36, 414 P.3d 1 (an appellate court may affirm a district court's ruling allowing evidence on any ground supported by the record) (cert. granted on other grounds Aug. 31, 2015).
¶ 15 "[E]xperts may not offer their direct opinion on a child victim's truthfulness or their opinion on whether children tend to fabricate sexual abuse allegations." People v. Wittrein , 221 P.3d 1076, 1081 (Colo. 2009). Expert testimony regarding typical behavior of abused children is admissible only if it addresses "general characteristics evidence which (1) relates to an issue apart from credibility and (2) only incidentally tends to corroborate a witness's testimony." People v. Cernazanu , 2015 COA 122, ¶ 20, 410 P.3d 603.
¶ 16 Although Ms. Tani didn't explicitly say that Y.P. was being "truthful," by opining that Y.P. didn't show any indication of having been coached, she conveyed the impression that she thought Y.P. was being truthful. Ordinarily, such testimony shouldn't be allowed. See People v. Eppens , 979 P.2d 14, 17-18 (Colo. 1999) (error to admit social worker's testimony that she felt the child sexual abuse victim "was sincere" in her forensic interview); People v. Snook , 745 P.2d 647, 649 (Colo. 1987) (social worker's testimony that children tend not to fabricate stories of sexual abuse was inadmissible because it went to the witness's truthfulness on a particular occasion); Cernazanu , ¶¶ 16-23 (district court erred in allowing mother's testimony that child victim didn't engage in her typical "lying" behavior when she reported sexual assault by the defendant; such testimony "necessarily implied" that mother thought the victim hadn't lied); People v. Bridges , 2014 COA 65, ¶ 16, 410 P.3d 512 (interviewer's testimony that child victims "had not been coached constituted conclusions about their truthfulness in their respective interviews" and was therefore inadmissible opinion).
¶ 17 We aren't persuaded to the contrary by the People's assertion that testifying that a child didn't show signs of having been coached is not the same as testifying that the child hasn't been coached, and therefore isn't an assessment of the child's credibility. The subtle distinction urged by the People is likely to be lost on ordinary jurors; rather, ordinary jurors, putting two and two together, are likely to glean from such testimony that the interviewer believed the child hadn't been coached.
¶ 18 Nor are we persuaded by the People's argument that Ms. Tani's testimony was admissible under CRE 608. While CRE 608(a) sometimes allows a witness to present evidence of another witness's character for truthfulness, it doesn't allow a lay or expert witness to testify that "another witness was telling the truth on a specific occasion ." Wittrein , 221 P.3d at 1081 (emphasis added); see also People v. Lopez , 129 P.3d 1061, 1066 (Colo. App. 2005) ("[A] witness's or prosecutor's *865personal opinion on the credibility of witnesses intrudes upon the province of the jury to make credibility determinations."). Because Ms. Tani testified to Y.P.'s truthfulness on a specific occasion (during her interview), and not generally, the testimony at issue wasn't admissible under CRE 608. See People v. Gaffney , 769 P.2d 1081, 1088 (Colo. 1989) (" CRE 608(a) would have permitted the prosecuting attorney to elicit on direct examination of [the witness] her opinion only as to [the child victim]'s general character for truthfulness, but not, as here, the [witness]'s opinion that [the child victim] was speaking the truth on a particular occasion.").
¶ 19 But that doesn't end our analysis concerning admissibility. For in this case, the record shows that defendant opened the door to the otherwise inadmissible testimony.
¶ 20 A party may open the door to otherwise inadmissible evidence by presenting incomplete evidence on a subject. "The concept of 'opening the door' represents an effort by courts to prevent one party in a criminal trial from gaining and maintaining an unfair advantage by the selective presentation of facts that, without being elaborated or placed in context, create an incorrect or misleading impression." People v. Murphy , 919 P.2d 191, 195 (Colo. 1996) (citing People v. Miller , 890 P.2d 84, 98-99 (Colo. 1995) ). Once a party opens the door, the opponent may inquire into the otherwise inadmissible matter. Golob v. People , 180 P.3d 1006, 1012-13 (Colo. 2008) (district court erred by allowing the People's expert to testify extensively on the defense expert's conclusions but then unduly limiting the defense expert's testimony on the subject).
¶ 21 As defendant essentially concedes,1 at trial he relied on theories that (1) Y.P.'s family members coached her on the details of the sexual assault and (2) Y.P. fabricated the allegation when she saw that her friend had received attention at school after reporting that she had been sexually abused.
¶ 22 Defense counsel initially presented these theories in opening statement:
[Y.P.'s friend] received attention in that school ... [a]nd so, lo and behold, they get this house just chuck full of people who are going to testify here, and they are all in that house together just waiting. And, of course, I'll ask them, you know, [w]ere you talking about the case? And I suspect they will say, [n]o, we weren't talking about the case, but they are all sitting in that house waiting for [the detective] to come down to talk to them.2
¶ 23 And defense counsel pursued these theories throughout trial. During cross-examination of Y.P., defense counsel asked her if she had practiced her testimony with the district attorney, if the women in her family liked to gossip about other family members, what "bad" things she had heard defendant had done, and whether she had talked with other family members about the incident just before she was interviewed. He also elicited Y.P.'s testimony that her cousin had reminded her of various details of the day of the incident. And he repeatedly underscored inconsistencies in Y.P.'s reports (including why the children had been called inside immediately prior to the assault, the nature and extent of the touching by defendant during the assault, and that none of the three cousins Y.P. testified to telling immediately after the assault could remember such a conversation).
*866He specifically asked each of Y.P.'s cousins if numerous family members were together in the house before the detective came to talk to them. All of this was clearly intended to suggest to the jurors that Y.P. had fabricated her allegations, at least partly as a result of coaching by family members and the district attorney. And all of this occurred before Ms. Tani testified.
¶ 24 Courts in other jurisdictions have held that by challenging a child witness's credibility by suggesting that the witness had been coached, the defense opens the door to testimony that the witness didn't appear to have been coached. Sampson v. State , 38 N.E.3d 985, 991-92 (Ind. 2015) (testimony by an interviewer as to whether a child witness showed signs of having been coached is admissible if the defendant opens the door by suggesting that the child had been coached); State v. Champagne , 371 Mont. 35, 305 P.3d 61, 67 (2013) (expert's testimony that she didn't see any signs that the child victim had been coached was admissible where the defendant had implied that the child had been coached); State v. Baymon , 336 N.C. 748, 446 S.E.2d 1, 4 (1994) ( "[D]efense counsel ... attempted to leave the impression that the victim had been coached by her relatives or social workers involved in the case. This attempt opened the door for the State on redirect to reestablish the reliability of the videotaped interview by proffering [the witness]'s testimony that she did not perceive that the victim had been told what to say or coached."); see also People v. Jefferson , 2014 COA 77M, ¶ 39, 411 P.3d 823 ("Unless defendant were to somehow 'open the door,' the prosecution's expert should not opine that a forensic interviewer's job is to determine whether or not a child is telling the truth."), aff'd , 2017 CO 35, 393 P.3d 493 ; cf. Venalonzo v. People , 2017 CO 9, ¶ 44, 388 P.3d 868 (by asking a detective whether, in his experience, children make things up, defense counsel opened the door to the detective's otherwise inadmissible testimony that, in his experience, children only make up trivial stories, not serious accusations). We agree with those courts' straightforward application of the opening the door principle. And so we conclude that defense counsel opened the door to Ms. Tani's testimony by suggesting that Y.P. had been coached.
¶ 25 In sum, though we disagree with the district court's reason for allowing Ms. Tani's testimony, we conclude that her testimony was nonetheless admissible. Therefore, the district court didn't abuse its discretion.
B. Prior Acts
¶ 26 Over defense counsel's objection, the district court allowed the prosecutor to introduce evidence of defendant's prior sexually related acts involving F.V. (Y.P.'s mother), E.V. (defendant's niece), and N.C. (also defendant's relative). The court told the jurors they could consider the evidence only as it served to refute a defense of recent fabrication or of impossibility; as it bore on defendant's mental state, knowledge, or intent; as it tended to show absence of mistake or accident; or as it tended to show defendant acted "for the purpose of sexual arousal, gratification, or abuse." Defendant contends that these other acts were too dissimilar to his alleged assault of Y.P. to be admissible under the applicable rule and statute. We don't agree.
1. Preservation, Applicable Law, and Standard of Review
¶ 27 The parties agree that defendant preserved this issue.
¶ 28 CRE 404(b) prohibits the admission of evidence about a defendant's prior acts when offered to show his bad character and that he acted in conformity with that character. Kaufman v. People , 202 P.3d 542, 552 (Colo. 2009). But other act evidence is admissible for other reasons, including to show "proof of motive, opportunity, intent, preparation, plan, knowledge, identity, or absence of mistake or accident." CRE 404(b) ; see also § 16-10-301(3), C.R.S. 2017 (in sexual offense prosecutions, the People may introduce other act evidence for similar purposes).
¶ 29 Before admitting evidence under CRE 404(b) or section 16-10-301, the court must determine that (1) the evidence relates to a material fact; (2) the evidence is logically relevant; (3) the logical relevance is independent of the intermediate inference that the defendant was acting in conformity with his bad character; and (4) the evidence has probative *867value that is not substantially outweighed by the danger of unfair prejudice. Kaufman , 202 P.3d at 552 ; People v. Snyder , 874 P.2d 1076, 1078 (Colo. 1994) ; People v. Spoto , 795 P.2d 1314, 1318 (Colo. 1990).3
¶ 30 Because the district court has substantial discretion in determining whether other act evidence is admissible, we won't overturn the court's decision unless it's manifestly arbitrary, unreasonable, or unfair. Kaufman , 202 P.3d at 553. And, because defendant preserved this issue, if we determine that the court abused its discretion, we review any error for harmless error. Yusem v. People , 210 P.3d 458, 469 & n.16 (Colo. 2009) ; People v. Munoz , 240 P.3d 311, 319 (Colo. App. 2009). An error is harmless unless there is a reasonable probability that it contributed to a defendant's conviction by substantially influencing the verdict or impairing the fairness of the trial. Yusem , 210 P.3d at 469 ; People v. Rincon , 140 P.3d 976, 979-80 (Colo. App. 2005).
2. Analysis
a. Acts Involving F.V. and N.C.
¶ 31 F.V. testified that when she was twenty-six years old, she was in a car with defendant when he touched her thigh and told her he "could take her to heaven" like her husband could not.
¶ 32 N.C. testified that she had lived with defendant briefly when she was nineteen years old. On multiple occasions, while family members were nearby, defendant approached her from behind, grabbed her, and tried to kiss her. She also testified that on one occasion defendant grabbed her from behind, touched her breasts and "below her stomach," and pressed his erect penis against her.
¶ 33 This evidence related to the material fact of whether defendant "knowingly subject[ed] another not his or her spouse to any sexual contact." § 18-3-405(1), C.R.S. 2017 (defining sexual assault on a child); see also Spoto , 795 P.2d at 1318 ; People v. Villa , 240 P.3d 343, 351 (Colo. App. 2009).
¶ 34 It also met the second prong of the Spoto test. Evidence is logically relevant if it has any tendency to make the existence of a material fact more or less probable than it would be without the evidence. CRE 401 ; see Yusem , 210 P.3d at 463. The evidence that defendant physically assaulted F.V. and N.C., two relatives who lived with him, made it more probable that defendant's intent was to sexually assault another female relative at his home. People v. Martinez , 36 P.3d 154, 159 (Colo. App. 2001) (other act evidence was logically relevant because "it had a tendency to make [the] defendant's intent to commit [the charged] crimes and the victim's lack of consent or lack of recent fabrication more probable with the evidence than without it"). And the other acts evidence was also relevant to rebut defendant's theory of defense that Y.P. fabricated the allegation. See id. ; see also People v. Orozco , 210 P.3d 472, 477-78 (Colo. App. 2009) ; People v. Victorian , 165 P.3d 890, 893 (Colo. App. 2007).
¶ 35 We are not persuaded by defendant's argument that F.V. and N.C.'s ages or sexual maturity render these prior acts too dissimilar to the act alleged in this case to make them relevant. " CRE 404(b) contains no separate requirement of similarity." People v. Rath , 44 P.3d 1033, 1041 (Colo. 2002). Although similarity may be necessary to give the evidence probative force if admitted for certain purposes, see id. at 1042, "it is not essential that the means of committing the other crimes replicate in all respects the manner in which the crime charged was committed." People v. Garner , 806 P.2d 366, 375 (Colo. 1991). In any event, the differences in the ages of the victims are offset by other similarities of the acts-specifically, that defendant sexually assaulted female family members, and did so in similar ways.4
¶ 36 We also conclude the acts have relevance aside from any inference of propensity. This prong of the Spoto test doesn't demand the absence of the propensity inference, but requires only that the evidence be logically relevant for a reason independent of that *868inference. Snyder , 874 P.2d at 1080 ; Villa , 240 P.3d at 352. As discussed, the evidence was probative of defendant's intent and his method of seeking sexual gratification from women to whom he had access because they were relatives who lived or stayed in his home. It was also probative to refute defendant's claim that Y.P. fabricated the allegation.
¶ 37 Lastly, we must consider whether the probative value of the evidence was substantially outweighed by the danger of unfair prejudice. In assessing the probative value of the evidence, we must evaluate the evidence's "incremental" probative value-what weight the evidence adds to the prosecution's case. Rath , 44 P.3d at 1041. In doing so, we weigh " 'the logical force of the evidence and the proponent's need for the evidence,' in light of other available evidence." Id. (quoting Martin v. People , 738 P.2d 789, 794 (Colo. 1987) ). We then must balance the evidence's incremental probative value against the danger of unfair prejudice, "afford[ing] the evidence the maximum probative value attributable by a reasonable fact finder and the minimum unfair prejudice to be reasonably expected." Id. at 1043.
¶ 38 Prior act evidence in sexual assault cases often has the potential for unfair prejudice. But we're mindful that the General Assembly has declared that "normally the probative value of [other sexual misconduct] evidence will outweigh any danger of unfair prejudice, even when incidents are remote from one another in time." § 16-10-301(1).
¶ 39 Given that Y.P.'s testimony was the only direct evidence of defendant's guilt, the other act evidence was especially relevant. See Rath , 44 P.3d at 1043 ("[W]here disputed testimony of the victim is the only direct evidence of the commission of the guilty act, additional evidence that is probative of that fact may have particular 'marginal' or 'incremental' probative value."). And while there was some potential for unfair prejudice, that potential didn't outweigh the evidence's substantial probative value. See People v. Mata , 56 P.3d 1169, 1173-74 (Colo. App. 2002) (probative value of evidence that the defendant digitally penetrated and fondled his daughter on multiple occasions, which was admitted to show intent and to refute a defense of fabrication by the victim, outweighed danger of unfair prejudice); People v. Duncan , 33 P.3d 1180, 1183-84 (Colo. App. 2001) (evidence that the defendant previously took three young men to secluded areas to have sexual contact with them would not "inflame the emotions of the jurors to reach an irrational decision").5
¶ 40 Therefore, we conclude that the district court didn't abuse its discretion by admitting the evidence of defendant's acts involving F.V. and N.C.
b. Acts Involving E.V.
¶ 41 The evidence regarding defendant's acts involving E.V. was somewhat different from that of the acts involving F.V. and N.C. E.V., defendant's niece, lived with him when she was seventeen years old. She testified that on four or five occasions, defendant masturbated while standing in the bedroom where she and defendant's two daughters were sleeping.
¶ 42 Unlike defendant's prior acts involving F.V. and N.C., he did not touch E.V. But his prior acts involving E.V. shared some similarities with the assault alleged by Y.P. Specifically, both involved female family members in his house, and E.V.'s age was closer to Y.P.'s than was F.V.'s or N.C.'s. On balance, the acts involving E.V. were as relevant as those involving F.V. and N.C., and the evidence of those acts was no more potentially prejudicial than the evidence of the acts involving F.V. and N.C. So this evidence likewise satisfied the four-part Spoto test.
¶ 43 But even if we assume the district court erred in allowing this evidence, we conclude that any error was harmless. The harmless error analysis requires "an inquiry into whether, viewing the evidence as a whole, the contested evidence substantially influenced the verdict or affected the fairness of the trial proceedings." People v. Summitt , 132 P.3d 320, 327 (Colo. 2006). We will disregard *869an evidentiary error if there is no "reasonable possibility that it contributed to the jury's guilty verdict." Id.
¶ 44 Because the district court properly admitted evidence of defendant's other prior acts involving family members F.V. and N.C., we conclude that there was no reasonable possibility that, had the evidence about E.V. been excluded, the outcome of the case would have been different.
III. Conclusion
¶ 45 The judgment is affirmed.
JUDGE FOX and JUDGE FREYRE concur.