JUSTICE BRUTINEL, opinion of the Court:

¶ 1 During Mark Haskie, Jr.'s trial on felony charges arising from an incident of domestic violence, Dr. Kathleen Ferraro, testifying as an expert witness, described general behavioral tendencies of adult victims of domestic abuse. Haskie argues that Dr. Ferraro's testimony should have been excluded as impermissible profile evidence. Because the testimony helped the jury understand the victim's behavior and was more probative than prejudicial, the trial court did not err in admitting it.

I. BACKGROUND

¶ 2 Haskie assaulted his girlfriend, P.J., at a Flagstaff motel after searching through messages on her phone and threatening her, "I told you I would kill you if you cheated on me." That same day, P.J. wrote a statement for the police explaining that Haskie had beaten and strangled her. Physical evidence from the motel corroborated her statement. Haskie was arrested nearly a year later. Shortly after his arrest, P.J. wrote two letters to the prosecutor recanting her earlier statements to the police, claiming instead that her injuries were from a bar fight she could not remember and that Haskie was innocent.

¶ 3 Before trial, the State filed a motion in limine to admit testimony by Dr. Ferraro as a "cold" expert on domestic violence to help the jury understand why P.J. had "continued her relationship with the defendant," "given conflicting statements while the case [was] pending," and why she was "reluctant to testify." The State's motion was accompanied by a list of questions the prosecutor intended to ask Dr. Ferraro. Haskie objected to Dr. Ferraro's proposed testimony, arguing it would not assist the jury and that it would constitute improper profile evidence and vouching. Following a hearing, the trial court limited Dr. Ferraro's testimony to the list of questions.

¶ 4 At trial, the State presented recorded phone calls Haskie made from jail, including several to P.J. before she recanted. In these conversations, Haskie dictated to P.J. an exculpatory story for her to tell police, apologized to her, and promised to marry her when he was released. During one call, P.J. responded, "[W]ell maybe you shouldn't have tried to kill me.... You know exactly what you did." At trial, however, P.J. testified that she did not remember who had beaten her because she had been drinking, and that although she initially blamed Haskie for her injuries because she was jealous, she had in fact cheated on him.

¶ 5 At trial, Dr. Ferraro testified that she was a "cold" or "blind" expert, meaning she had not reviewed any case-specific evidence and was not going to testify about any of the events in the case. The prosecutor asked her a series of questions regarding characteristics of domestic violence victims to help the jury understand behaviors that might otherwise seem counterintuitive to jurors unfamiliar with domestic violence. When asked, "[I]s it unusual for someone who has been hurt by an intimate partner to return to that relationship?" Dr. Ferraro responded, "It's not unusual. It is very common." She continued, "There are many reasons [why,] and they vary by the individual, of course, and the type of relationship." Dr. Ferraro explained that some victims of domestic violence return to their abusers out of fear, retaliation, or threats, while others do not leave their abusers because of pressure from extended family or the victim's own shame. Dr. Ferraro further testified that chemical dependency and alcohol abuse complicate the decision to leave an abusive relationship.

¶ 6 The prosecutor then asked, "[D]o victims ever tend to blame themselves for what happened?" Dr. Ferraro responded:

The prosecutor also asked, "Is it unusual for victims to later change their story?" Dr. Ferraro answered, "No, that is very typical," adding that victims recant or change the details of their account for many of the reasons that might also make a victim reluctant to leave the relationship. In addition, she explained, the victim may be afraid of violent repercussions; may feel pressure from the abuser or friends and extended family; may be intimidated to discontinue prosecution; and may be emotionally and psychologically manipulated.

¶ 7 Then the following exchange took place:

¶ 8 During closing arguments, the prosecutor did not mention Dr. Ferraro or compare any aspect of her testimony to the facts of Haskie's case. Before jury deliberations began, the trial court instructed the jurors that they were not bound by any expert opinion and should give an opinion only the weight they believed it deserved. The jury found Haskie guilty of two counts of aggravated assault (domestic violence), five counts of aggravated domestic violence, two counts of influencing a witness, and one count of kidnapping.

¶ 9 The court of appeals affirmed, holding that Dr. Ferraro's testimony did not constitute impermissible profile evidence. State v. Haskie, 240 Ariz. 269, 273 ¶ 18, 276 ¶ 34, 378 P.3d 446 (App. 2016).

¶ 10 We granted review to consider whether Dr. Ferraro's testimony constituted impermissible offender profiling.1 We have jurisdiction pursuant to article 6, section 5(3) of the Arizona Constitution and A.R.S. § 12-120.24.

II. DISCUSSION

¶ 11 We review a trial court's admission of evidence for an abuse of discretion, which can include errors of law. State v. Ketchner, 236 Ariz. 262, 264 ¶ 13, 339 P.3d 645 (2014); see also State v. Cheatham, 240 Ariz. 1, 2 ¶ 6, 375 P.3d 66 (2016). We review interpretation of court rules de novo. State v. Salazar-Mercado, 234 Ariz. 590, 592 ¶ 4, 325 P.3d 996 (2014).

¶ 12 Initially, we note that in Salazar-Mercado, a case involving child victims, we held that Arizona Rule of Evidence 702 permits the admission of "cold" expert testimony that educates the fact-finder about general principles without applying those principles to the particular facts of the case. 234 Ariz. at 591 ¶ 1, 325 P.3d 996. Salazar-Mercado's rationale applies equally to cases involving adult victims.

¶ 13 In Ketchner we precluded testimony that "implicitly invited the jury to infer criminal conduct based on the [cold expert's descriptions of] characteristics," relying on Ryan v. State, 988 P.2d 46, 56-57 (Wyo. 1999). 236 Ariz. at 265 ¶¶ 17, 19, 339 P.3d 645 (holding that the cold expert's testimony constituted impermissible profile evidence). We now elaborate.

¶ 14 "Profile evidence tends to show that a defendant possesses one or more of an informal compilation of characteristics or an abstract of characteristics typically displayed by persons engaged in a particular kind of activity." Id. at 264 ¶ 15, 339 P.3d 645 (internal quotation marks and citations omitted). Describing evidence as "profile" evidence is a shorthand way of saying that the evidence is offered to implicitly or explicitly suggest that because the defendant has those characteristics, a jury should conclude that the defendant must have committed the crime charged.

¶ 15 The state may not offer "profile" evidence as substantive proof of the defendant's guilt. See id. at 264-65 ¶¶ 15-19, 339 P.3d 645. The rationale for this rule is evident: "[P]rofile evidence may not be used as substantive proof of guilt because of the `risk that a defendant will be convicted not for what he did but for what others are doing.'" Id. ¶ 15 (quoting State v. Lee, 191 Ariz. 542, 545 ¶ 12, 959 P.2d 799 (1998)).

¶ 16 Conversely, expert testimony that explains a victim's seemingly inconsistent behavior is admissible to aid jurors in evaluating the victim's credibility. See State v. Moran, 151 Ariz. 378, 381, 728 P.2d 248 (1986) (citing State v. Lindsey, 149 Ariz. 472, 474, 720 P.2d 73 (1986)). Although expert testimony about victim behavior that also describes or refers to a perpetrator's characteristics has the potential to be "profile" evidence, it is not categorically inadmissible. Rather, its admissibility is determined by the rules of evidence. The burden of establishing admissibility lies with the proponent of the testimony — in this case, the State. Like all evidence, such testimony must be relevant to be admissible. See Ariz. R. Evid. 401 ("Evidence is relevant if: (a) it has any tendency to make a fact more or less probable than it would be without the evidence; and (b) the fact is of consequence in determining the action."); Ariz. R. Evid. 402 ("Relevant evidence is admissible unless any of the following provides otherwise: the United States or Arizona Constitution; an applicable statute; these rules; or other rules prescribed by the Supreme Court. Irrelevant evidence is not admissible.").

¶ 17 If relevant, such evidence may still be excluded if the prejudice created by its admission substantially outweighs its probative value. See Ariz. R. Evid. 403 ("The court may exclude relevant evidence if its probative value is substantially outweighed by a danger of one or more of the following: unfair prejudice, confusing the issues, misleading the jury, undue delay, wasting time, or needlessly presenting cumulative evidence."). Thus, evidence of offender characteristics may be admissible, subject to a Rule 403 analysis, if it is relevant for a reason other than to suggest that the defendant possesses some of those characteristics and therefore may have committed the charged crimes. See Lee, 191 Ariz. at 546 ¶ 19, 959 P.2d 799 ("[T]here may be situations in which drug courier profile evidence has significance beyond the mere suggestion that because an accused's conduct is similar to that of other proven violators, he too must be guilty.").

¶ 18 The outcome of this analysis will, of course, vary from case to case. "Deciding whether expert testimony will aid the jury and balancing the usefulness of expert testimony against the danger of unfair prejudice are generally fact-bound inquiries uniquely within the competence of the trial court." Moran, 151 Ariz. at 381, 728 P.2d 248. The more "general" the proffered testimony, the more likely it will be admissible. See State v. Chapple, 135 Ariz. 281, 292, 660 P.2d 1208 (1983), superseded by statute on other grounds as stated in State v. Goudeau, 239 Ariz. 421, 459 ¶ 154, 372 P.3d 945 (2016). In addition, the more the testimony is tied to the defendant's characteristics, rather than to those of the victim, the more likely the admission of such testimony will be impermissibly prejudicial. See id.

¶ 19 The danger of "cold" evidence describing the interaction between offenders and victims is that it may stray into prejudicial and potentially improper profile evidence. Ketchner provides an example of such evidence. The expert witness (also Dr. Ferraro) in Ketchner testified about "characteristics common to domestic violence victims and their abusers," and "described risk factors for `lethality' in an abusive relationship." 236 Ariz. at 264 ¶ 14, 339 P.3d 645. This Court held the testimony was inadmissible because, in addition to explaining victim behavior that otherwise might be misunderstood by a jury, it described an abuser's reaction to loss of control in a relationship, inviting a comparison with the defendant's actions. Id. at 265 ¶ 19, 339 P.3d 645. Importantly, in Ketchner, the victim's actions were not at issue and the expert's testimony did not explain victim behavior: "There was no reason to elicit this testimony except to invite the jury to find that Ketchner's character matched that of a domestic abuser who intended to kill or otherwise harm his partner in reaction to a loss of control over the relationship." Id. In other words, the expert testimony in Ketchner was simply not relevant to explaining the victim's behavior.

¶ 20 Here, the victim's behavior and inconsistent statements were squarely at issue. Dr. Ferraro's testimony was limited to questions designed to help the jury understand the sometimes counterintuitive behaviors of domestic violence victims. Although a few of her general statements referred to an abuser's characteristics, such as, "the abusive partner will turn the violence around and say that if you hadn't done this or you had done that as I told you to do, this never would have happened, so it's your fault," each statement primarily served the purpose of explaining victim behavior. Thus, the testimony was relevant to help the jury understand P.J.'s behavior.

¶ 21 The trial court considered the admissibility of Dr. Ferraro's testimony at the hearing on the State's motion in limine. At that hearing, Haskie neither objected to Dr. Ferraro's testimony on Rule 403 grounds nor argued prejudice. The trial court found that the testimony was relevant and granted the motion.

¶ 22 Based on this record, the trial court did not abuse its discretion in admitting Dr. Ferraro's testimony. Any prejudice from her testimony was minimal and did not outweigh the testimony's probative value. The evidence revealed some characteristics of domestic violence abusers mentioned by Dr. Ferraro. As the court of appeals recognized, "[I]t is not surprising — indeed it is expected — that the jury will hear evidence that the victim has behaved to a greater or lesser extent in accord with the testimony of a `cold' and `blind' expert such as Dr. Ferraro." Haskie, 240 Ariz. at 274 ¶ 23, 378 P.3d 446. Our case law has recognized that "just because expert testimony about behavioral characteristics is exceedingly persuasive does not mean it is unfairly prejudicial." Moran, 151 Ariz. at 384, 728 P.2d 248 (emphasis in original) (holding that the trial judge did not abuse his discretion in admitting such testimony under Rules 403 and 702). Dr. Ferraro's testimony was not directed at establishing that Haskie possessed "one or more of an informal compilation of characteristics" typically displayed by domestic violence abusers; rather, it was introduced to explain the impetus for the victim's counterintuitive behavior. Ketchner, 236 Ariz. at 264 ¶ 15, 339 P.3d 645. She neither explicitly nor implicitly invited the jury to infer criminal conduct based on the described conduct. See id. at 265 ¶ 17, 339 P.3d 645 (citing with approval Ryan, 988 P.2d at 55 (warning that even testimony that only implicitly invites the jury to infer criminal conduct on the part of the defendant based on described characteristics demands close scrutiny under the character evidence rules)). Dr. Ferraro never made comparisons between general characteristics of an abusive relationship and the facts of this case. During closing arguments, the prosecutor never mentioned Dr. Ferraro or related any aspect of her testimony to P.J. or Haskie.

¶ 23 The State proposes a standard for identifying when profile evidence is impermissibly prejudicial. It argues that to violate the prohibition against using profile evidence as substantive proof of guilt, the prosecution must offer testimony (1) establishing the existence of a common profile for perpetrators of a certain criminal activity, (2) enumerating the profile's component characteristics, and (3) expressly comparing the defendant against each component characteristic to establish guilt by showing that he "matches" the profile in most or all respects. But requiring an explicit mention of a "profile" or a direct comparison of the defendant with the expert's testimony leaves too much room for prejudice and ignores the real possibility that an expert could create a profile without ever explicitly describing it as such.

¶ 24 We conclude that the trial court should consider the prejudicial effect of the expert's testimony as a whole, as well as that of each individual statement offered. See State v. Steinle, 239 Ariz. 415, 419 ¶ 14, 372 P.3d 939 (2016) (stating that Rule 403 issues "are highly contextual — they necessarily depend on assessments of not only the evidence in question, but also the other evidence in the case"). But piecing together statements, none of which make direct comparisons to the defendant's conduct and all of which are relevant to explaining a victim's behavior, does not necessarily establish that the evidence is more prejudicial than probative.

¶ 25 Although admission of Dr. Ferraro's testimony in this case was not error, we note that trial courts should exercise great caution in screening, admitting, and limiting this type of evidence. Evidence describing the characteristics of offenders, even as part of a description of victim behavior, could imply that a defendant is guilty. This potential for undue prejudice requires that trial courts carefully scrutinize such evidence.

¶ 26 If such testimony is admitted, the defendant is entitled to a limiting instruction under Rule 105 of the Arizona Rules of Evidence to explain to the jury the limited purpose and scope of such testimony. Cf. Woodson v. State, 30 Ariz. 448, 455, 247 P. 1103 (1926) ("[W]hen evidence [that when considered as a whole is highly prejudicial and is bound to engender hostility toward the defendant] is admitted it is the duty of the trial court to use extraordinary care in instructing the jury and in seeing that no improper matter is allowed to get before it."). Additionally, although testimony about offender or victim characteristics from a "cold," "blind" expert is not categorically inadmissible, that does not mean a trial court should automatically admit it. Rather, trial courts should filter such proffered evidence through the screens of Rules 401, 402, 403, and 702 of the Arizona Rules of Evidence. Furthermore, we caution trial courts to limit "cold" and "blind" testimony from expert witnesses to matters within the scope of their expertise. Such experts should not be allowed to speak in broad, categorical terms about supposedly "common" or "usual" occurrences without empirical support.

III. CONCLUSION

¶ 27 We vacate paragraphs seventeen through twenty-three of the court of appeals' decision and affirm Haskie's convictions and sentences.