On November 16, 1984, a jury convicted William Moran (defendant) of one count of child molestation and two counts of sexual abuse. A.R.S. §§ 13-1404, 13-1410. The crimes allegedly were committed on his daughter. He was sentenced to fourteen years for child molestation and two and one-half years for each sexual abuse count. The court of appeals affirmed. State v. Moran, 151 Ariz. 373, 728 P.2d 243 (App.1985).

At trial, defendant’s daughter recanted her allegations that she had been sexually abused by her father. There was no physical evidence of abuse. Consequently, the only evidence linking defendant with the alleged abuse was testimony of witnesses recounting the daughter’s out-of-court reports of the molestation. The credibility of the daughter’s out-of-court statements was bolstered by several experts testifying that the statements were truthful and that the daughter’s behavior, including recantation, was typical of molested children. We accepted review to address the admissibility of the expert testimony. Rule 31.19, Ariz.R.Crim.P., 17 A.R.S. (Supp.1985). We have jurisdiction under Ariz. Const, art. 6, § 5(3) and A.R.S. § 12-120.24.

FACTS AND ISSUES

Defendant has been married for seventeen years and, in addition to the alleged victim, has two sons. Defendant’s daughter first reported the alleged sexual abuse to her high school principal. She claimed that her father had been having sexual contact with her since she was five years old. Sheriff’s deputies investigating the complaint recorded the daughter’s statement. Defendant was arrested and[*380] charged with child molestation and sexual abuse.

At trial, under oath, the daughter recanted her accusations. However, she admitted having told school authorities, two detectives, two therapists, her mother, her foster mother, her friend, and her friend’s mother that she had been sexually abused by her father. Seven witnesses testified about the daughter’s extra-judicial accusations. The jury also heard the daughter’s tape-recorded statement. No other direct or physical evidence was offered to establish the charged offenses.

Defendant argues that the trial court erred in admitting the daughter’s pri- or inconsistent statements as the only evidence of guilt. In a well-reasoned opinion, the court of appeals held the statements admissible under Rule 801(d)(1)(A), Ariz.R.Evid., 17A A.R.S. ^[1] The court distinguished State v. Allred, 134 Ariz. 274, 655 P.2d 1326 (1982), and concluded that the prior inconsistent statements could be used as substantive evidence of the crime, even if they were the only evidence of the crime. 151 Ariz. at 375-376, 728 P.2d at 245-246. We agree with the court of appeals’ analysis on this issue.

Defendant also argues that the trial court erred in allowing expert witnesses to testify (1) that they believed the daughter was telling the truth when she first reported the crime and (2) that the daughter’s behavioral characteristics “matched” characteristics of other child victims of sexual abuse. Defense counsel moved before trial to preclude this type of testimony, and objected throughout trial to specific testimony about both credibility and common behavior characteristics. The trial court denied defendant’s pretrial motion and all subsequent objections.

The court of appeals held the expert testimony admissible, discounting the danger that the jury would overvalue the experts’ “obvious belief in the truthfulness of the daughter’s” out-of-court statements. At 377, 728 P.2d at 247. The court concluded that the trial court was correct in not excluding the expert testimony merely because of a “risk of overweighting.” Id. Although' we agree with much of the court’s reasoning, we do not agree that all the expert testimony in this case was admissible.

EXPERT TESTIMONY IN SEXUAL ABUSE CASES

A. General Principles

We recently considered the propriety of expert testimony in sexual abuse cases in State v. Lindsey, 149 Ariz. 472, 720 P.2d 73 (1986). Lindsey held that admissibility of expert testimony in sexual abuse cases is governed by the same rules of evidence applicable to all expert opinion testimony. ^[2] Id. at 475, 720 P.2d at 76. Under Rules 702, 703, and 403, expert testimony must (1) come from a qualified expert, (2) be reliable, (3) aid the triers of fact in evaluating and understanding matters not within their common experience, and (4) have probative value that outweighs its prejudicial[*381] effect. State v. Chapple, 135 Ariz. 281, 291, 660 P.2d 1208, 1218 (1983).

Defendant did not object to the experts’ qualifications or the reliability of their testimony. Therefore, as in Lindsey, we must answer two questions: first, did the expert testimony provide the jurors with useful information outside their common understanding or experience, Rule 702; and second, was the usefulness of the expert testimony “substantially outweighed by the danger of unfair prejudice,” Rule 403.

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. State v. Neal, 143 Ariz. 93, 100, 692 P.2d 272, 279 (1984); State v. Mincey, 141 Ariz. 425, 441, 687 P.2d 1180, 1196, cert. denied, 469 U.S. 1040, 105 S.Ct. 521, 83 L.Ed.2d 409 (1984); State v. Williams, 132 Ariz. 153, 160, 644 P.2d 889, 896 (1982). However, when the admissibility of expert opinion evidence is a question of “law or logic,” it is this court’s responsibility to determine admissibility. Chappie, 135 Ariz. at 297 n. 18, 660 P.2d at 1224 n. 18. For example, Lindsey held that certain types of expert opinion evidence will not assist juries in sexual abuse cases and therefore are inadmissible under Rule 702. 149 Ariz. at 475, 720 P.2d at 76. If we determine that certain types of opinion testimony are inherently unhelpful, then any danger of prejudice requires holding the evidence inadmissible. Id.; see also Abrams v. Interco, 719 F.2d 23, 28 (2d Cir.1983) (appellate court need not “honor a purported exercise of discretion which was infected by an error of law”).

B. Types of Expert Testimony

1. Testimony About General Behavioral Characteristics

Lindsey recognized that expert testimony on recantation and other problems afflicting sexual abuse victims may explain a victim’s seemingly inconsistent behavior and aid jurors in evaluating the victim's credibility. 149 Ariz. at 474, 720 P.2d at 75. Other jurisdictions, recognizing the usefulness of expert testimony in child sexual abuse cases, also allow experts to explain general behavioral characteristics of child victims. E.g., People v. Dunnahoo, 152 Cal.App.3d 561, 577, 199 Cal.Rptr. 796, 804 (Cal.Ct.App.1984) (testimony explaining delay in reporting); Smith v. State, 100 Nev. 570, 571-72, 688 P.2d 326, 326-27 (1984) (same); People v. Benjamin R., 103 A.D.2d 663, 668-69, 481 N.Y.S.2d 827, 831-32 (N.Y.App.Div.1984) (same); State v. Middleton, 294 Or. 427, 436-37, 657 P.2d 1215, 1220 (1983) (recantation, truancy, and tendency to run away from home); Commonwealth v. Baldwin, 348 Pa.Super. 368, 372-73, 502 A.2d 253, 255 (Pa.Super.Ct. 1985) (reporting delays and inconsistent versions of abuse); State v. Petrich, 101 Wash.2d 566, 575-76, 683 P.2d 173, 179-80 (1984) (delay in reporting). Oregon Supreme Court Justice Roberts explained the rationale for allowing this type of expert testimony:

While jurors may be capable of personalizing the emotions of victims of physical assault generally, and of assessing witness credibility accordingly, tensions unique to the trauma experienced by a child sexually abused by a family member have remained largely unknown to the public. As the expert’s testimony demonstrates the routine indicia of witness reliability — consistency, willingness to aid the prosecution, straight forward rendition of the facts — may, for good reason, be lacking. As a result jurors may impose standards of normalcy on child victim/witnesses who consistently respond in distinctly abnormal fashion.

Middleton, 294 Or. at 440, 657 P.2d at 1222 (Roberts, J., concurring).

We agree with Justice Roberts’s analysis. “Jurors, most of whom are unfamiliar with the behavioral sciences, may benefit from expert testimony” explaining behavior they might otherwise “attribute to inaccuracy or prevarication.” Lindsey, 149 Ariz. at 474, 720 P.2d at 75; accord State v. Myers, 359 N.W.2d 604, 610 (Minn.1984) (allowing expert testimony explaining “puzzling aspects of the child’s conduct and[*382] demeanor which the jury could not otherwise bring to its evaluation of [the victim’s] credibility”). Such evidence may harm defendant’s interests, but we cannot say it is unfairly prejudicial; it merely informs jurors that commonly held assumptions are not necessarily accurate and allows them to fairly judge credibility. See State v. Chapple, supra.

2. Particularized Testimony About the Alleged Victim’s Credibility

Although Lindsey allowed expert testimony explaining child victims’ seemingly strange behavior, the court placed strict limits on more particularized opinion evidence in sexual abuse cases. We held that expert testimony may not be admitted “to ‘tell the jury’ who is correct or incorrect, who is lying and who is truthful.” 149 Ariz. at 474, 720 P.2d at 75; accord United States v. Azure, 801 F.2d 336, 339 (8th Cir.1986) (expert testimony that victim was “believable” improperly admitted under Rule 702); Kruse v. State, 483 So.2d 1383, 1387-88 (Fla.Dist.Ct.App.1986); Middleton, 294 Or. at 438, 657 P.2d at 1221. Nor may the expert's opinion as to credibility be adduced indirectly by allowing the expert to quantify the percentage of victims who are truthful in their initial reports despite subsequent recantation. Lindsey, 149 Ariz. at 474, 720 P.2d at 75; accord State v. Myers, 382 N.W.2d 91, 97 (Iowa 1986) (improper to admit expert testimony that children rarely lie about sexual abuse).

The rationale behind Lindsey’s limitation on expert testimony is that once the emotional antecedents underlying the victim’s behavior are explained, “the jury needs nothing further from the expert.” 149 Ariz. at 475, 720 P.2d at 76; cf. State v. Hicks, 133 Ariz. 64, 71, 649 P.2d 267, 274 (1982) (expert testimony on effects of alcohol intoxication not allowed because jury has adequate knowledge; therefore, jury must decide defendant’s intent the same way a psychiatrist would — “by looking at the facts surrounding the shooting”). Some courts reason that the marginal benefit of expert testimony on the question of guilt or innocence is insufficient to outweigh the substantial prejudice such evidence is likely to plant in the minds of jurors. E.g., Petrich, 101 Wash.2d at 576, 683 P.2d at 180. The danger of prejudice may be heightened when, as here, the state’s case rests entirely on extra-judicial statements.

We do not consider admission of particularized testimony about credibility a Rule 403 balancing situation. We see no reason to risk influencing the jury’s credibility determination by allowing expert opinion testimony on a witness’s believability. ^[3] With proper information, the jury can evaluate credibility as well as an expert. We believe that the expert’s conclusory opinion is of no assistance; therefore, Rule 702 precludes admission. Azure, 801 F.2d at 339-41 (Rule 702 prohibits direct opinion testimony that child victim of sexual abuse is telling the truth); see also United States v. Barnard, 490 F.2d 907 (9th Cir.1973) (trial court properly rejected defendant’s psychiatric testimony that government’s witness was a lying sociopath), cert. denied, 416 U.S. 959, 94 S.Ct. 1976, 40 L.Ed.2d 310 (1974); State v. Williams, 132 Ariz. 153, 160, 644 P.2d 889, 896 (1982). An expert’s belief in a witness’s credibility “has never been a permissible subject of expert opinion less the trial process return to the discredited notion of marshalling adherents of either side as oathtakers.” M. UDALL & J. LIVERMORE, LAW OF EVIDENCE § 22, at 30-31 (2d ed. 1982).

Further, in cases that hinge on the credibility of a witness’s allegations of sex[*383] ual abuse, a psychologist’s opinion that one witness, and not another, should be believed infringes upon the jury’s prerogative to determine the ultimate question in the case. Azure, 801 F.2d at 340 (“ ‘It is horn-book law that the credibility of a witness and the weight to be given his testimony rests exclusively with the jury.’ ”) (quoting United States v. Rosenburg, 108 F.Supp. 798, 806 (S.D.N.Y.), aff'd, 200 F.2d 666 (2d Cir.1952)). We are aware that Rule 704 permits “testimony in the form of an opinion or inference” even though such testimony “embraces an ultimate issue to be decided” by the jury. Rule 704.. However, Rule 704 was not intended to permit experts to tell “the jury what result to reach.” Fed. R.Evid. 704, advisory committee note, reprinted in 56 F.R.D. 183, 284 (1973). When the only evidence consists of the victim’s accusation and the defendant’s denial, expert testimony on the question of who to believe is nothing more than advice to jurors on how to decide the case. Such testimony was not legitimized by Rule 704, and is not admissible under Rule 702. The same principle applies to expert opinion testimony on whether the crime occurred, whether the defendant is the perpetrator, and like questions. ^[4] See, e.g., Petrich, 101 Wash.2d at 576, 683 P.2d at 180 (where defendant was related to victim, statement that in “ ‘eighty-five to ninety percent of our cases, the child is molested by someone they already know,’ ” should not have been admitted).

C. Application to Facts

At trial, the defense asserted that the daughter’s initial accusation was prompted by anger at her parents. The defense explained that the victim had been disciplined for misbehaving at school, and therefore had wanted to leave home and live with a friend. Defendant argued that the daughter’s anger and school problems stemmed from parental discipline, not parental abuse. The daughter subsequently recanted, the defense argued, because she realized she was unable to live with her friend, wanted to go home, and knew it was wrong to have made a false charge. The defense repeatedly argued that it was incomprehensible that the daughter would want to return home if her father had molested her.

Most of the prosecution’s expert testimony was aimed at explaining that the daughter’s behavior, particularly her recantation, was not inconsistent with abuse having occurred. Several experts explained that anger is a typical response to sexual molestation and that even the daughter’s problems at school may have been caused by abuse. Similarly, experts explained factors that could lead a victim to recant and attempt to return home. ^[5]

[*384] This type of expert testimony was properly admitted. Defendant claimed that the victim’s accusations were prompted by anger over discipline imposed by her parents. Testimony providing the jury with an alternative explanation for the victim’s anger was admissible to assist the jury in determining what had motivated the initial charges against defendant. Similarly, evidence explaining why recantation is not necessarily inconsistent with the crime having occurred aided the jury in evaluating the victim’s credibility. Testimony describing behavioral characteristics or conduct outside jurors’ common experience is permitted under the principles articulated in Chappie and Lindsey as long as it meets other requirements of admissibility. Rule 702.

Obviously, even general testimony may unduly prejudice the jury. Consequently, the trial judge has discretion to exclude general testimony if he or she finds that the benefits are substantially outweighed by the potential for unfair prejudice. State v. Chappie, supra; Rule 403. Here, however, several considerations support the conclusion that ¿the testimony at issue was properly admitted. First, just because expert testimony about behavioral characteristics is exceedingly persuasive does not mean it is unfairly prejudicial. Cf United States v. Jimenez, 789 F.2d 167, 171 (2d Cir.1986). Second, the defense attorney can, and in this case did, effectively cross-examine the experts by pointing out facts the experts failed to consider. For example, at least three experts admitted that although the literature describes recantation as typical, none of their patients, other than the victim in this case, had ever recanted. Third, most of the expert testimony on behavioral characteristics was fairly general; it described and explained the seemingly strange behavior displayed by many child victims without opining as to this victim’s veracity. When this type of testimony is involved, we agree with the court of appeals that the “good common sense of jurors will discern that which is true from that which is false.” 151 Ariz. at 377, 728 P.2d at 247. We hold that the trial judge did not abuse his discretion in admitting such testimony under Rules 403 and 702.

However, some of the expert testimony in this case went far beyond describing general principles of behavior. One expert testified as follows, giving a nonresponsive answer to a question about the significance of the daughter’s reports of frequent sexual contact:

A: My personal opinion is that the only thing [the daughter] has lied about is the extent. I believe there was penetration. [The daughter had denied penetration.]

Q: [Y]ou believe in fact that there had been some penetration?

A: Yes. ^[6]

T.R., 11/9 at 183-84 (footnote added). The same witness testified that the daughter was referred to her by Child Protective Services “to help this young lady deal with the fact that she had been molested and to deal with the other issues that were resulting from that, low self esteem, depression, being out of the home.” T.R., 11/9 at 146 (emphasis added).

Obviously, this testimony was intended to tell the jury that the experts[*385] believed the daughter’s earlier version of the abuse and that she had been molested. This type of testimony is prohibited by Lindsey. Experts called to testify about behavioral characteristics that may affect an alleged victim’s credibility may not give an opinion of the credibility of a particular witness. Psychologists and psychiatrists are not, and do not claim to be, experts at discerning truth. Psychiatrists are trained to accept facts provided by their patients, not to act as judges of patients’ credibility. People v. Bledsoe, 36 Cal.3d 236, 250, 681 P.2d 291, 300, 203 Cal.Rptr. 450, 459 (1984); see also Azure, 801 F.2d at 341 (“no reliable test for truthfulness exists”). ^[7]

Another psychologist testified that Child Protective Services asked her “to do an evaluation to ascertain whether or not / felt [the daughter] has been sexually mo lested____” T.R., 11/15 at 10 (emphasis added). The affirmative results of that evaluation were implicit in the balance of the witness’s testimony. The same psychologist gave the daughter a variety of personality tests. She then testified as follows:

Q: Doctor, overall were the findings that you got with regard to [the daughter] on these objective personality ^[8] tests consistent with an individual who had in fact some kind of trauma like a molest occur early in life and were now simply reacting to that in adolescence?

A: Yes, it is consistent.

T.R., 11/15 at 41 (footnote added). ^[9]

This type of testimony — indicating that the victim’s behavior is consistent with the crime having occurred — is slightly different than direct testimony on the victim’s veracity. However, like the testimony disallowed in Lindsey and the expert’s assertion in this case that penetration occurred, the inference offered the jury is that because this victim’s personality and behavior are consistent with a molest having occurred, the crime must have been committed. See State v. Huey, 145 Ariz. 59, 63, 699 P.2d 1290, 1294 (1985) (we “might have some difficulty in upholding the admissibility of rape trauma syndrome to prove the existence of rape”).

This type of particularized testimony permits the expert to indicate how he or she views the credibility of a particular witness. Once the jury has learned the victim’s behavior from the evidence and has heard experts explain why sexual abuse may cause delayed reporting, inconsistency, or recantation, we do not believe the jury needs an expert to explain that the victim’s behavior is consistent or inconsistent with the crime having occurred. See Azure, 801 F.2d at 340-41 (testimony about general behavioral characteristics may assist the jury, but specific testimony regarding the victim’s credibility is impermissible); People v. Roscoe, 168 Cal.App.3d 1093, 1099-1100, 215 Cal.Rptr. 45, 49-50 (Cal.Ct.App.1985) (general explanation of[*386] sexual abuse victims’ behavior allowed; specific testimony regarding expert’s examination of current victim not allowed); cf State v. Chapple, 135 Ariz. at 292, 660 P.2d at 1219 (the “generality” of expert testimony on the dynamics of eyewitness identification is a factor favoring admission); Fuenning v. Superior Court, 139 Ariz. 590, 605, 680 P.2d 121, 136 (1983) (in DWI case, police officer may testify about symptoms exhibited by defendant that indicate intoxication, but may not opine that defendant was intoxicated).

In some cases, eliciting an expert’s opinion that the victim’s conduct is consistent with the crime having occurred might be harmless. However, there is always a significant risk that such testimony will allow the expert to convey his or her belief that this victim was molested and is telling the truth or has recanted under pressure. We believe this happened here. Given the other expert testimony available to the jury, we believe the expert’s testimony that the victim’s behavior and personality were consistent with the crime having occurred did not assist the jury. It was error to allow such testimony. Rules 702 and 704.

CONCLUSION

We hold that the trial court should not have admitted testimony that the victim’s behavior was consistent with the abuse having occurred. Further, the court erred in permitting an expert to imply her belief of the daughter’s veracity and in permitting the expert’s “personal opinion” that the daughter was telling the truth about the molestation and lying only about the extent of penetration. Such testimony was inadmissible under Rules 702 and 704.

Neither physical evidence nor any other direct evidence showed that Moran committed the crime. The only evidence was the out-of-court statements, later recanted at trial. Therefore, we must hold that the errors, taken together, were prejudicial.

The judgment of conviction is reversed and the case is remanded for a new trial. The opinion of the court of appeals is approved in part and vacated in part.