JUSTICE BOATRIGHT
delivered the Opinion of the Court.
¶1 Petitioner Julio Venalonzo was convicted of sexual assault on a child, attempted sexual assault on a child, possession of drug paraphernalia, and resisting arrest. Venalon-zo appealed, and the court of appeals affirmed in an unpublished opinion. People v. Venalonzo, No. 07CA0882, slip op. at 1, 2011 WL 4837489 (Colo. App. Oct. 13, 2011). Here, we determine whether a forensic interviewer who testified as a lay witness crossed the line between lay and expert testimony. We also address whether the interviewer, the mother of one of the child victims, and an investigating police officer improperly testified that the child victims’ accusations against Venalonzo were truthful.1
¶2 First, we hold that in determining whether testimony is lay testimony under Colorado Rule of Evidence (“CRE”) 701 or expert testimony under CRE 702, the trial [871]*871court must look to the basis for the opinion. If the witness provides testimony that could be expected to be based on an ordinary person’s experiences or knowledge, then the witness is offering lay testimony. If, on the other hand, the witness provides testimony that could not be offered without specialized experiences, knowledge, or training, then the witness is offering expert testimony. Applying that holding, we conclude that some portions of the interviewer’s testimony in this case were admissible as lay opinion but that others were inadmissible expert testimony in the guise of lay opinion. Second, we hold that the interviewer’s and the mother’s testimony improperly bolstered the credibility of the child victims by creating an impermissible inference that they were telling the truth in this case. However, we hold that Venalonzo opened the door to the investigating officer’s statements regarding what children lie about, and therefore the trial court did not abuse its discretion in admitting those statements. Finally, we conclude that the errors in this case warrant reversal. Accordingly, we reverse the court of appeals’ judgment and remand this case to that court with instructions to return the case to the trial court for further proceedings consistent with this opinion.2
I. Facts and Procedural History
¶3 Seven-year-old A.M. and eight-year-old C.O. told their respective parents that a man called them over to him while the children were playing in the hallway of their apartment complex and then touched A.M. sexually. Later that day, the two girls both made statements to the police, and, as a result of those statements, police arrested Venalonzo. During the course of the investigation, A.M. and C.O. also spoke with a child forensic interviewer, Ann Smith (“the interviewer”), at Sungate Children’s Advocacy Center (“the Advocacy Center”). Both children also testified at trial.
¶4 The children never wavered in them account that a man called them over to him, and A.M. never wavered in her account that a man inappropriately touched her. There were, however, differences between A.M.’s and C.O.’s accounts. Specific details that varied included what the man who assaulted A.M. was wearing, what he was doing with his hands before he touched A.M., where he was standing, what he said to the children, how and from where he entered the building, to whom the children first told their story, and whether C.O. was present when the man assaulted A.M. Each child’s own account also changed over time on these details. Moreover, A.M.’s story changed regarding where the man had touched her, including whether it was over or under her clothing, and what he did before and after.
¶5 Before trial, Venalonzo requested a summary of any expert or “specialized knowledge” testimony. The People maintained that they would not present any expert testimony, so the trial court denied Ven-alonzo’s request. Prior to and during trial, Venalonzo objected to portions of the interviewer’s testimony that he asserted were expert testimony. The trial court deferred its ruling on the objections that Venalonzo made prior to trial until it heard the interviewer’s testimony. At trial, the interviewer began by testifying about the interviewing process that she used, her training, and her experience interviewing children. Venalonzo objected to this background, arguing that it called for expert testimony. The trial court called a bench conference, during which Venalonzo’s defense counsel renewed his objection to the interviewer’s anticipated testimony about A.M.’s and C.O.’s behavior during the interview. He argued that relating and interpreting the demeanor and language of the children during the forensic interviews — as well as discussing whether the children’s behavior was common or uncommon as compared with other victims of sexual abuse — was expert testimony and that such testimony amounted to statements that the children were telling the truth in this case.
¶6 After the bench conference, the court declined to require the People to qualify the interviewer as an expert and ruled that she [872]*872would be allowed to answer the People’s questions as long as she did not comment on the children’s veracity. The court did, however, allow the interviewer to testify about whether A.M.’s and C.O.’s behavior during their interviews was common among children whom she previously interviewed.
¶7 After this ruling, the People continued them direct examination. Venalonzo then specifically objected that the People were attempting to elicit expert testimony that bolstered A.M.’s and C.O.’s credibility when the People asked the interviewer to provide the following information: (1) how many children the interviewer had forensically interviewed with the Advocacy Center; (2) how many children the interviewer ultimately determined had not been abused; (3) how many times the interviewer had testified in a court of law; (4) whether children commonly tell other children about abuse before telling adults; (5) whether children typically say different things to the interviewer than they do to responding officers or Department of Human Services workers, or when testifying at trial; (6) whether specific details on which the children’s stories differed — such as whether C.O. was present when A.M. was assaulted— were core or peripheral; (7) what “reproduction” — whereby child victims demonstrate inappropriate touching on their own bodies — is; (8) how A,M. “reproduced” the event in question; (9) why forensic interviewers look for signs of reproduction; and (10) whether forensic interviews sometimes lead to charges being dropped. The trial court allowed the interviewer to answer all of these questions except for two: the question regarding the gestures A.M. used to reproduce the event, ruling that this evidence was cumulative because the jury would witness A.M.’s actions when it watched the videotaped interview, and the question regarding why interviewers look for signs of reproduction, ruling that this called for expert testimony.
¶8 Venalonzo also objected to the testimony from two other witnesses: A.M.’s mother (“Mother”) and the investigating police officer. The People called Mother to testify about A.M.’s outcry. During her testimony, she said that A.M. had never accused anyone else of touching her inappropriately and that she did not know of any reason for A.M. to make up such a story, stating further, “She wouldn’t unless it was true.” To support this position, the prosecutor asked her to testify about the signs that reveal when her daughter is lying. Mother did so, and she testified that A.M. had exhibited none of those signs when reporting the alleged assault that gave rise to this ease. Finally, she stated that A.M.’s mind “wouldn’t go to [the] level” of falsely accusing someone of sexual assault.
¶9 The investigating police officer also testified. On direct examination, he told the jury about his interviews with A.M. and C.O. on the day of the incident. During cross-examination, the officer agreed with defense counsel that, based on his prior experience as a school teacher, young children were suggestible and sometimes made up stories or could be talked into doing so by other children. On redirect, the prosecutor asked the officer to clarify what types of stories children tended to make up. Venalonzo objected, asserting that the testimony would constitute improper bolstering, but the trial court ruled that Ven-alonzo had opened the door to the question. The officer responded that children make up stories about “trivial things” but that he had never experienced children of the victim’s age “making things up ... about something of a very serious nature.”
¶10 Ultimately, the jury returned guilty verdicts on both the sexual assault and the attempted sexual assault charges, and the trial court sentenced Venalonzo to an indeterminate prison sentence of three years to life for sexual assault on a child and a consecutive two-year prison sentence for attempted sexual assault on a child. Venalonzo appealed, arguing that the trial court erred when it admitted the interviewer’s testimony because (1) the People had not endorsed her as an expert witness and (2) her testimony amounted to an opinion that A.M. and 0.0. were telling the truth. Additionally, Venalonzo argued that the trial court erred when it allowed Mother and the investigating police officer to offer what amounted to testimony that A.M. and C.O. were telling the truth.
¶11 The court of appeals affirmed the trial court’s rulings on all issues. The court held that the interviewer’s testimony about her [873]*873qualifications and training was basic information within the scope of lay opinion testimony. Venalonzo, slip op. at 9-10. Relying on People v. Tillery, 231 P.3d 36 (Colo. App. 2009), the court also noted that the interviewer’s testimony comparing the children’s interviews in this case to typical child victim interviews “was derived from her ‘years of observing such interviews,’ not.from specialized training.” Venalonzo, slip op. at 10 (quoting Tillery, 231 P.3d at 42). The court of appeals then held that “the interviewer gave no opinion on whether A.M. and C.O. told the truth in their interviews,” and it stated that “[t]he likelihood that the jury could infer from the interviewer’s testimony that the girls’ accounts were credible does not mean the interviewer expressed that assessment.” Id. at 15. Rather, the court held that the interviewer’s testimony was admissible because it “spoke to the general characteristics of child victim interviews, which helped the jury assess the girls’ credibility.” Id.
¶12 The court of appeals concluded further that the interviewer’s testimony was helpful because “it informed the jury that the inconsistencies in children’s relation of events are common.” Id. It noted that Venalonzo had opened the door to this testimony in his opening statement by suggesting that the children’s inconsistent stories meant that they were lying, and that the interviewer’s testimony properly responded to that attack on their credibility. Id Finally, it held that the interviewer properly testified that forensic interviews sometimes result in exonerating suspects because the interviewer was responding to Venalonzo’s implication during cross-examination that the interviewer and the organization she worked for were biased in favor of law enforcement. Id. at 16.
¶13 The court of appeals also rejected Venalonzo’s arguments regarding Mother and the investigating police officer. As to Mother, it held that her testimony went to AM.’s sophistication and general characteristics for truthfulness. In doing so, the court of appeals stated that trial courts “may admit opinion testimony about the general credibility of a child sex assault victim if it will ‘assist the jury in evaluating the credibility of the testifying child [victim]’ and if it relates to general characteristics for truthfulness.” Id. at 11-12 (quoting People v. Woertman, 786 P.2d 443, 447 (Colo. App. 1989), rev'd on other grounds, 804 P.2d 188 (Colo. 1991)). It also noted that Mother never explicitly testified that A.M. was telling the truth and that, in any event, Venalonzo had opened the door by implying on cross-examination that AM. had been following C.O.’s lead when she made up the allegations. As to the officer, it held that his testimony regarding what types of things children make up stories about was proper. According to the court of appeals, Venalonzo opened the door to this testimony, and it was relevant because it “helped the jury ‘make a more informed decision in evaluating the credibility of a testifying child.’” Id. at 14 (quoting People v. Koon, 713 P.2d 410, 411 (Colo. App. 1985)). The court of appeals rejected all of Venalonzo’s remaining arguments.
¶14 We granted Venalonzo’s petition for certiorari. Ultimately, we reverse the convictions.
II. Standard of Review
¶15 We review a trial court’s eviden-tiary decisions for an abuse of discretion. People v. Stewart, 55 P.3d 107, 122 (Colo.2002). A trial court abuses its discretion when its ruling is manifestly arbitrary, unreasonable, or unfair. Id.
III. Analysis
¶16 We first consider Venalonzo’s contention that the trial court erred in admitting the interviewer’s testimony without requiring that the People endorse her as an expert witness. We hold that in determining whether testimony is lay testimony under CRE 701 or expert testimony under CRE 702, the trial court must look to the basis for the opinion. If the witness provides testimony that could be expected to be based on an ordinary person’s experiences or knowledge, then the witness is offering lay testimony. If, on the other hand, the witness provides testimony that could not be offered without specialized experiences, knowledge, or training, then the witness is offering expert testimony. Applying that holding, we conclude that some portions of the interviewer’s testimony in this [874]*874case were admissible as lay opinion but that others were inadmissible expert testimony in the guise of lay opinion. Second, we consider Venalonzo’s argument that the interviewer's, Mother’s, and the investigating officer’s testimony violated the rule against testimony that a witness is telling the truth in a specific instance. We hold that the interviewer’s -and Mother’s testimony improperly bolstered A.M.’s and C.O.’s credibility by creating an impermissible inference that they were telling the truth in this case. However, we hold that Venalonzo opened the door to the investigating officer’s statements regarding what children lie about, and therefore the trial court did not abuse its discretion in admitting those statements. Finally, we conclude that the errors in this case warrant reversal.
A. Expert Testimony
1. Law
¶17 Whether a witness’s testimony is lay or expert depends on the facts and surrounding circumstances of the case and “requires a ease-by-case analysis of both the witness and the witness’s opinion.” United States v. Smith, 591 F.3d 974, 982-83 (8th Cir. 2010). To make this determination, we must start with the Colorado Rules of Evidence.
¶18 Together, CRE 701 and 702 distinguish lay testimony from expert testimony. CRE 701 defines the scope of lay witness opinion testimony. It provides that lay witness testimony in the form of opinions or inferences must be “(a) rationally based on the perception of the witness, (b) helpful to a clear understanding of the witness’[s] testimony or the determination of a fact in issue, and (e) not based on scientific, technical, or other specialized knowledge within the scope of [CRE] 702.” CRE 702, on the other hand, concerns the admissibility of expert testimony. Under this rule, “[i]f scientific, technical, or other specialized knowledge will assist the 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.” CRE 702.
¶19 Previously, we considered this distinction between lay and expert testimony - in Stewart. There, a police officer testified about his training in investigating and reconstructing traffic accidents, his observations and reconstruction of the crime scene at issue in the case, and the inferences he drew from the reconstruction. Stewart, 55 P.3d at 122-23. We held that the officer’s testimony about his observations of the crime scene and his investigation of the incident were admissible as lay opinion testimony. Id. at 124. His “deductions about ... the vehicle’s direction, position, and speed” during the accident, however, required specialized training and knowledge and were therefore expert testimony. Id. Accordingly, we concluded that the trial court abused its discretion in admitting the investigating police officer’s accident reconstruction testimony without requiring that he be qualified as an expert. Id.
¶20 Similarly, in People v. Veren, 140 P.3d 131, 139-40 (Colo. App. 2005), the court of appeals considered whether a police officer’s testimony that he suspected a defendant of manufacturing methamphetamine solely because the defendant possessed a large amount of Sudafed was expert testimony. The court held that this was improper expert testimony in the guise of lay opinion because it required specialized knowledge. Id. at 139. The court of appeals reasoned that while ordinary people might know that Sudafed contains an ingredient that can be used to manufacture methamphetamine, they probably would not know how much Sudafed would be required for this purpose. Id. The court of appeals noted that “the only persons having such knowledge would be those who are either actually involved in the sale of- illegal substances, or those who are involved in law enforcement’s effort to curb such sales.” Id. (quoting State v. Rothlisberger, 95 P.3d 1193, 1200 n.5 (Utah App. 2004)). Therefore, the Veren court held that the officer’s testimony was expert testimony because, without specialized knowledge and training, the officer would not have been able to infer based on the facts that he observed that the defendant intended to manufacture and sell methamphetamine. Id.
[875]*875¶21 While this court has addressed the line between lay and expert testimony in the context of police officers, we have not done so for forensic interviewers. Thus, case law from other jurisdictions is informative. State v. Gonzalez, 150 N.H. 74, 834 A.2d 354 (2003), is particularly relevant to our inquiry here. In Gonzalez, the New Hampshire Supreme Court addressed a social worker’s testimony that victim recantations or denials are not unusual in sexual abuse cases. Id. at 358. The prosecutor first asked the social worker whether she had received any training or education about sexual abuse allegations, and then asked whether recantations are unusual in sexual abuse cases. Id at 356. The social worker responded that recantations are not unusual. Id The court held that the witness’s statements concerning “a child’s delayed disclosure of abuse, inconsistent statements about abuse, and recantation of statements about abuse" constituted expert testimony because those behaviors “may be puzzling or appear counterintuitive to lay observers when they consider the suffering endured by a child who is continually being abused.” Id. at 358 (quoting State v. Cressey, 137 N.H. 402, 628 A.2d 696, 702 (1993)). In other words, the testimony was improper because an ordinary person would not know that children commonly delay disclosing abuse.
¶22 As these cases illustrate, the critical factor in distinguishing between lay and expert testimony is the basis for the witness’s opinion. That is, the proper inquiry is not whether a -witness draws on her personal experiences to inform her testimony; all witnesses rely on their personal experience when testifying. Rather, it is the nature of the experiences that could form the opinion’s basis that determines whether the testimony is lay or expert opinion. See Stewart, 55 P.3d at 123; see also People v. Souva, 141 P.3d 845, 849 (Colo. App. 2005) (concluding that an eye witness who happens to be a certified addictions counselor may offer lay testimony about whether the defendant appeared to be under the influence of drugs at the time she observed him). To determine whether the testimony in question is testimony that an ordinary person could give, “courts consider whether ordinary citizens can be expected to know certain information or to have had certain experiences.” People v. Rincon, 140 P.3d 976, 982 (Colo. App. 2005) (citing United States v. McDonald, 933 F.2d 1519, 1522 (10th Cir. 1991)). Expert testimony, by contrast, is that which goes beyond the realm of common experience and requires experience, skills, or knowledge that the ordinary person would not have. See James River Ins. Co. v. Rapid Funding, LLC, 658 F.3d 1207, 1214 (10th Cir. 2011).
¶23 Hence, we hold that in determining whether testimony is lay testimony under CRE 701 or expert testimony under CRE 702, the trial court must look to the basis'for the opinion. If the witness provides testimony that could be expected to be based on an ordinary person’s experiences or knowledge, then the witness is offering lay testimony. If, on the other hand, the witness provides' testimony that could not be offered without specialized experiences, knowledge, or training, then the witness is offering expert testimony.
¶24 This distinction can be a difficult one, and because it requires trial courts to engage in fact-speeific analyses in the course of trial, we review their conclusions for an abuse of discretion.
¶25 With this understanding in mind, we now turn to the trial court’s decision to admit the interviewer’s testimony as lay opinion in this ease.
2. Application
¶26 Here, Venalonzo challenges specific portions of the interviewer’s testimony. We treat these challenges separately because we conclude that some, but not all, of her testimony exceeded the scope of CRE 701.
¶27 First, we address Venalonzo’s argument that the interviewer’s statements about her training, experience, and techniques conducting forensic interviews constituted expert testimony.3 The interviewer’s [876]*876testimony describing her professional background, including the number of interviews she has conducted and the number of times she has testified in court, is not expert testimony because any ordinary person is capable of describing her own credentials. The interviewer also described how she conducts an interview with a child victim, including that she builds rapport with the child, tests the child’s memory, directs the child to answer questions honestly, and usually asks open-ended, not leading, questions. This description of basic interview protocol did not amount to expert testimony, either. While an ordinary person may not be able to articulate a specific protocol for conducting interviews, she would know that in order to obtain accurate information from a child, it is important to make the child feel comfortable, evaluate whether the child remembers the events in question, and make sure the child understands the importance of the truth. As for the distinction between leading and non-leading questions, the terms themselves may not be familiar to a lay person, but the concepts certainly are. The ordinary person could be expected to understand that asking a question, especially of a child, that in itself suggests the answer would yield a less reliable response. Significantly, in testifying about her professional background, the interviewer did not draw any inferences about the victim’s behavior in this case, and she did not describe her interview techniques in detailed or technical terms that would have been outside the understanding of an ordinary person. We conclude that this general description of the protocol that the interviewer uses when interviewing child victims did not require specialized knowledge, and thus the trial court did not abuse its discretion in admitting it.
¶28 Next, Venalonzo argues that the interviewer’s testimony that children are not very good at understanding physical measurements, that they often use generalities when speaking, and that they often reveal secrets to othér children before they tell adults was expert testimony. We disagree. That the interviewer recognized this behavior because of her years of experience interviewing child sex assault victims does not mean that her observations were expert testimony. The ordinary person has spent time with children and could reasonably be expected to know that they are not as accurate or perceptive as adults. Similarly, an ordinary person could be expected to know that children are more apt to share information with their peers than with adults, especially if they are unsure whether they may have done something wrong and fear being punished. Because an ordinary person who interacts with children can recognize these behaviors without additional training or specialized experience, this information is lay opinion testimony within the scope of CRE 701.
¶29 Although the aforementioned parts of the interviewer's testimony were not expert testimony, other portions that Venal-onzo challenges fall outside the scope of CRE 701 because they include conclusions that relied on the interviewer’s specialized experience. In particular, the interviewer testified as an expert when she described the significance of the children’s behavior during the forensic interviews. This included the interviewer’s reference to “reproduction” — the habit of children to gesture with their hands to demonstrate where on their bodies they were touched — and her statement that A.M. had engaged in this behavior in this case. She used her specialized knowledge to answer the prosecutor’s question about what a child’s reproduction means to a forensic investigation. She explained that reproduction is “[w]hen [child victims] actually demonstrate on their own body what occurred to them, they do that without my prompting, try to describe some touching that happened to them. They physically demonstrate that on themselves.” The interviewer could not have offered this testimony without her specialized experience conducting forensic interviews. While this testimony was brief, the interviewer’s explanation attached significance to the victims’ behavior that an ordinary person would not recognize, and its admission as lay testimony was thus an abuse of the trial court’s discretion,
[877]*877¶30 Similarly, the interviewer’s discussion of “core versus peripheral” details also included opinion based on specialized knowledge, particularly when she discussed the relative importance of each type of detail in determining the truth of an allegation and categorized several of the details to which the children testified as either core or peripheral.4 The significance of these differences is not common knowledge, nor would the ordinary person recognize their importance based on her life experience. Assigning weight to “core” and “peripheral” details is technical and requires training, skilled observation, and specialized knowledge. An ordinary person could not be expected to properly characterize parts of a child’s story as a core or peripheral fact so as to determine that story’s truth. While anyone who has made a credibility determination knows that the truth is found in the details of the story, giving special weight to “core details” is different. Some people may recognize the distinction based on their personal interactions with children, but the ordinary person would probably not know that the core details of a child's account will remain the same even if the child forgets or confuses the peripheral details. Here, the interviewer testified that she learned about the significance of core versus peripheral details from her extensive training in forensic interviewing and from her experience treating child sex assault victims. Thus, making this distinction constituted expert testimony that was “based on ... specialized knowledge within the scope of Rule 702,” meaning the trial court abused its discretion in admitting it as lay testimony. See CRE 701(c).
¶31 Having determined that it was error to permit the forensic interviewer to offer expert testimony in the guise of lay opinion, we now address whether three witnesses’ testimony improperly bolstered A.M.’s and C.O.’s credibility.
B. Bolstering Testimony
¶32 This court has held that witnesses are prohibited from testifying that another witness is telling the truth on a particular occasion. People v. Wittrein, 221 P.3d 1076, 1081 (Colo. 2009) (citing CRE 608(a)); People v. Eppens, 979 P.2d 14, 17 (Colo. 1999). This rule applies to both direct and indirect implications of a child’s truthfulness. See Wittrein, 221 P.3d at 1082 (citing People v. Snook, 745 P.2d 647, 648 (Colo. 1987), for the proposition that a social worker’s testimony that children tend not to fabricate stories of sexual abuse was inadmissible because it was tantamount to testimony that the child victim was telling the truth in that ease).5 The danger in admitting such testimony lies in the possibility that it will improperly invade the province of the fact-finder. See People v. Newbrough, 803 P.2d 155, 163 (Colo. 1990).
¶33 Testimony that another witness is credible is especially problematic where the outcome of the case turns on that witness’s credibility. This often occurs' in child sex assault cases. See, e.g., Pennsylvania v. Rit-[878]*878chie, 480 U.S. 39, 60, 107 S.Ct. 989, 94 L.Ed.2d 40 (1987) (“Child abuse is one of the most difficult crimes to detect and prosecute, in large part because there often are no witnesses except the victim”). Children may delay reporting sexual abuse, making the collection of physical evidence difficult or impossible. See People v. Fasy, 829 P.2d 1314, 1317 (Colo. 1992) (discussing expert testimony that child victims delay reporting sexual assault). In such instances, the child’s testimony is likely the most significant evidence, and the case turns on whether the fact-finder finds the alleged victim credible. See, e.g., State v. Marrington, 335 Or. 555, 73 P.3d 911, 917 (2003) (holding that the erroneous admission of expert testimony was reversible because the case “involved a swearing contest” in which the victim claimed that the defendant touched her, and he testified in his own defense that he did not). Thus, courts must be particularly mindful of testimony that a child victim is telling the truth when that child’s testimony is “the focal issue in the case.” Snook, 745 P.2d at 649.
¶34 With these principles in mind, we now assess whether the interviewer’s, Mother’s, and the officer’s testimony violated the rule against testimony that another witness is telling the truth.
2. Forensic Interviewer
¶36 As discussed above, the interviewer testified about the two girls’ interviews and compared their behavior to that of other child sex assault victims. Specifically, she stated that many of the children’s behaviors were common to other child sex assault victims she had interviewed and testified that some forensic interviews have led the People to drop charges against suspects. We conclude that the interviewer’s testimony violated ORE 608(a) because it improperly bolstered the children’s credibility and led to the impermissible inference that the children were telling the truth about the incident.
¶36 First, the only purpose for the interviewer’s testimony comparing A.M.’s and C.O.’s behavior to that of other child sex assault victims was to bolster the children’s credibility. See Snook, 745 P.2d at 648-49 (holding that “the jury’s only conceivable use of’ testimony that children tend not to fabricate stories of sexual abuse and that they often reproduce their experience when reporting it “would be as support for the complainant’s truthful character”). Admitting this evidence did not make any other fact at issue more or less probable. The testimony was not necessary for the People to lay a foundation for admitting the videotaped forensic interviews. Cf. CRE 901(b)(1) (providing that “[t]estimony that a matter is what it is claimed to be” is sufficient to lay foundation for admissibility). Once the interviewer began comparing A.M. and C.O. to other child sex assault victims — stating that children who had been sexually assaulted commonly gave conflicting details — her testimony had no proper purpose. Instead, the People offered it to show that, despite their inconsistent accounts of the assault, A.M. and C.O. were telling the truth. We conclude that the trial court abused its discretion in admitting this testimony.
¶37 Venalonzo also objected to the interviewer’s statement on redirect examination that forensic interviews have led the People to drop charges against some suspects. He argued that this testimony implied that A.M. and C.O. were telling the truth because, after the interviewer’s forensic interviews in this case, the People moved forward with their charges against Venalonzo. The People counter that this testimony was proper because they elicited the statement in response to Venalonzo’s implication that the interviewer was biased. We agree with Ven-alonzo. The prosecutor first brought up the issue of bias on his direct examination of the interviewer. Specifically, the prosecutor asked, “[A]re you an agent for the police?” to which the interviewer responded that she was not. Subsequently, during cross-examination, defense counsel pursued this line of questioning by asking which organizations hire the Advocacy Center to conduct forensic interviews and about the Advocacy Center’s relationship with various state governmental and legal entities. Although these questions implied bias, they did not open the door to the interviewer’s testimony that previous, unrelated interviews resulted in dropped charges. The People sufficiently responded to [879]*879Venalonzo’s implication by eliciting testimony from the interviewer that she was not biased when she interviewed AM. and C.O., that no district attorneys served on the Advocacy Center’s board or influenced its policies, and that forensic interviewers try to prevent suggestibility to protect the integrity of the case and the defendant’s interest. The testimony about previous interviews resulting in dropped charges implied that the children were telling the truth about Venalonzo’s acts because, here, the People charged Venalonzo with sexual assault. Thus, we conclude that the trial court abused its discretion in admitting it.6
¶38 We now address the portions of Mother’s testimony that Venalonzo challenges.
3. Mother
¶39 Mother testified that A.M. did not display any signs that she was lying when she reported the incident, that A.M. was not sophisticated enough to make up a story about the sexual assault, and that A.M. had no reason to accuse Venalonzo unless the incident had actually occurred. We conclude that Mother’s testimony amounted to testimony that AM. was telling the truth about the sexual assault.
¶40 Mother’s statements were similar to those that we held were improper in Snook and Wittrein. The expert social worker in Snook testified that “children tend not to fabricate stories of sexual abuse and in giving reports tend to reproduce their experiences.” 745 P.2d at 648. She further testified that, “in order to make these things up, there has to be a basis for that experience.” Id. We held that the testimony was inadmissible under CRE 608 because it was testimony “that [the child victim] is almost certainly telling the truth.” Id. at 648-49. Similarly, in Wittrein, a child psychiatrist testified, “It’s'hard for me to imagine that an eight-year-old child would be able to put together ... a plan [to portray herself as a victim].” 221 P.3d at 1081. We characterized this statement as an improper “generalization about whether, children have the sophistication to fabricate allegations of sexual abuse.” Id. at 1082. Thus, we held that the testimony was inadmissible. Id.
¶41 Here, although Mother did not directly state that her daughter was telling the truth, her statements communicated this belief. As in Snook and Wittrein, Mother’s testimony on this issue served no other purpose than to bolster A.M.’s credibility. By testifying that her daughter was not sophisticated enough to fabricate her allegations against Venalonzo, that she had no reason to make up the story, and that she would not make this accusation unless it were true, Mother expressed her belief that A.M. was telling the truth in this instance. Therefore, Mother’s testimony violated CRE 608(a), and the trial court abused its discretion by allowing it.
¶42 The People assert that Venalonzo opened the door to this testimony by implying that A.M. made up her story at C.O.’s suggestion. Even assuming, without deciding, that defense counsel’s questions attacked A.M.’s character for truthfulness, Mother’s testimony addressed A.M.’s truthfulness in this specific instance as opposed to her general character for truthfulness. When questioning Mother, the prosecutor stated, “you said that she wouldn’t accuse somebody of this,” and asked whether “her mind wasn’t sophisticated enough to come up with this?” (Emphasis added). He also asked Mother whether she knew of any reason A.M. would say that Venalonzo had touched her. The prosecutor’s questions thus elicited Mother’s comments on A.M.’s veracity in this case, specifically regarding Venalonzo. For this reason, even if Venalonzo opened the door to testimony about A.M.’s general character for truthfulness, Mother’s testimony went be[880]*880yond the scope of admissible testimony under CRE 608(a).
¶43 Next, we consider the investigating officer’s testimony.
4. The Investigating Officer
¶44 Venalonzo argues that the trial court erred in permitting the investigating police officer to testify that, in his experience as a school teacher, children only make up trivial stories, not serious accusations.7 Normally, this statement would constitute improper testimony that the children were telling the truth. See Snook, 745 P.2d at 648-49; Wittrein, 221 P.3d at 1081-82. In this case, however, Venalonzo had previously questioned the officer about this exact issue when he asked whether, in the officer’s experience, kids make things up. By doing so he opened the door for further questioning on this matter. The People were entitled to follow defense counsel’s question about kids making things up with a question regarding what types of things they would make up. “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.” Golob v. People, 180 P.3d 1006, 1012 (Colo. 2008). The officer’s testimony on redirect examination placed his answer on cross-examination in context and did not exceed the scope of cross-examination. Therefore, because Venalonzo opened the door to the officer’s statements, the trial court did not err in admitting this testimony.
¶45 In sum, although the trial court did not err in admitting the investigating officer’s statements because Venalonzo opened the door to this testimony, it did abuse its discretion in admitting portions of the interviewer’s and Mother’s testimony because they violated the rule against testimony that another witness was telling the truth.
¶46 Having reached these conclusions, we now turn to the question of whether the trial court’s errors in admitting the interviewer’s and Mother’s testimony merit reversal.
C. Reversible Error
¶47 Venalonzo contends that the trial court’s admission of the interviewer’s and Mother’s improper testimony prejudiced him because their testimony vouched for the two child witnesses’ credibility — which was a central issue in this case. We agree and therefore conclude that the trial court’s error in admitting the testimony was not harmless.
¶48 We review preserved, non-constitutional errors for harmless error. Hagos v. People, 2012 CO 63, ¶ 12, 288 P.3d 116, 119.8 Under harmless error review, we reverse only if the error “affects the substantial rights of the parties.” Id. An error affects a party’s substantial rights if it “substantially influenced the verdict or affected the fairness of the trial proceedings.” Id. (quoting Tevlin v. People, 715 P.2d 338, 342 (Colo. 1986)).
¶49 We considered the impact of a similar error in Snook. In that case, we held that the trial court’s erroneous admission of a social worker’s expert testimony that “children tend not to fabricate stories of sexual abuse” required reversal of the defendant’s conviction. Snook, 745 P.2d at 648-49. We reasoned that the victim’s credibility was the focal issue in the case, and because the social worker’s testimony “directly supported [the victim’s] credibility, it may have been the deciding factor in the jury’s decision that [the victim’s] version was correct.” Id. at 649. Like in Snook, here both the interviewer’s and Mother’s improper testimony substantially impacted Venalonzo’s right to a fair trial.
[881]*881¶50 As to the interviewer’s testimony, the prosecution’s failure to disclose the interviewer as an expert witness prejudiced Ven-alonzo because the interviewer’s specialized experience, combined with her use of technical terms, imbued her testimony with an air of expertise and may have led the jury to credit her assessment of the children’s credibility over other evidence in the ease. If Venalonzo had had the benefit of pretrial disclosure of the interviewer’s expert testimony and the bases for her opinions, then he would have had “the opportunity to evaluate the testimony in advance of trial or to obtain his own expert witness.” Veren, 140 P.3d at 140; accord Todd v. Bear Valley Vill. Apartments, 980 P.2d 973, 979 (Colo. 1999) (discussing similar prejudice in civil cases). Instead, when the interviewer offered an expert opinion on the most crucial evidence in the case, Venalonzo was without recourse to counter that testimony with his own expert witness. Therefore, permitting the interviewer to offer her expert opinion caused Venalonzo to suffer the prejudice that the rules governing expert witnesses seek to prevent. Furthermore, A.M.’s and C.O.’s testimony was the focal point of the ease, since no other witnesses were present during the alleged assault and there was no physical evidence. Because of the significance of A.M.’s and C.O.’s testimony, the interviewer’s improper expert testimony that directly supported them credibility by justifying inconsistencies in their testimony caused significant unfair prejudice. See Veren, 140 P.3d at 140.
¶51 Similarly, Mother’s testimony directly communicated to the jury that she believed A.M. was telling the truth. The prosecutor even pointed out in his closing argument that Mother is the one person who knows A.M. best, thus furthering the impact of her improper testimony.
¶52 In sum, we cannot say that the interviewer’s and Mother’s improper testimony did not substantially influence the jury’s verdict. Therefore, the trial court's errors in permitting the interviewer to offer improper expert testimony and in permitting the interviewer and Mother to offer testimony that improperly bolstered the children’s credibility were not harmless, and they require us to reverse Venalonzo’s convictions. See id.
IV. Conclusion
¶53 We hold that in determining whether testimony is lay testimony under CRE 701 or expert testimony under CRE 702, the trial court must look to the basis for the opinion. If the witness provides testimony that could be expected to be based on an ordinary person’s experiences or knowledge, then the witness is offering lay testimony. If, on the other hand, the witness provides testimony that could not be offered without specialized experiences, knowledge, or training, then the witness is offering expert testimony. Applying that holding, we conclude that some portions of the interviewer’s testimony in this case were admissible as lay opinion but that others were inadmissible expert testimony in the guise of lay opinion. We further hold that the interviewer’s and Mother’s testimony improperly bolstered A.M.’s and C.O.’s credibility by creating an impermissible inference that they were telling the truth in this ease. However, we hold that Venalonzo opened the door to the investigating officer’s statements regarding what children lie about, and therefore the trial court did not abuse its discretion in admitting those statements. Finally, we conclude that the errors in this case warrant reversal. Accordingly, we reverse the court of appeals’ judgment and remand this case to that court with instructions. to return the case to the trial court for further proceedings consistent with this opinion.
JUSTICE COATS concurs in the judgment, and JUSTICE EID joins in the concurrence in the judgment.