OPINION

SINGLETON, Judge.

Richard and Betty Anderson, who are husband and wife, were jointly charged, tried and convicted of multiple counts of sexual abuse of a minor in the first degree, an unclassified felony, AS 11.41.434(a)(1), and sexual abuse of a minor in the second degree, a class A felony, AS 11.41.436(a)(2). They appeal their convictions on many grounds. We reverse.

I. FACTS

J.E. and E.E. received day care from Betty Anderson in her home from August 1981 until July 1984. In the fall of 1983, their mother began to notice what she considered to be strange behavior by the children. In July of 1984, having heard rumors about the Andersons, their mother confronted J.E., age four and one-half, and E.E., age three and one-half, about possible sexual abuse. The children informed their mother that Betty and Richard Anderson had, in fact, sexually abused them. The mother reported the accusations to the day-care licensing bureau and a police officer. Shortly thereafter, she took the children to see Dr. Linda Ekman, a pediatrician. Dr. Ekman examined the children and noticed that each had anal tears consistent with sexual penetration of the anus. A grand jury was convened to hear these allegations. The state initially presented a four-count indictment which, at the request of the grand jury, was expanded to eleven counts. The grand jury returned an eleven-count indictment.

II. DISCUSSION A. Mental Health Examination of Child Abuse Victims

The Andersons argue that the trial court erred in denying their request to have the children examined by a court-appointed mental health practitioner in order to determine the children's competency and to aid the defense in evaluating their credibility. In appropriate cases, trial courts clearly have discretion to order such mental health examinations. Braham v. State, 571 P.2d 631, 639-40 (Alaska 1977), cert. denied, 436 U.S. 910, 98 S.Ct. 2246, 56 L.Ed.2d 410 (1978); Pickens v. State, 675 P.2d 665, 668-69 (Alaska App. 1984).

In order to avoid victim harassment, we have adopted a two-prong balancing test to reconcile the defendant's right to gather facts relevant to his defense with the victim's constitutionally protected right to privacy. First, the defendant seeking a psychiatric examination of a prosecuting witness must make a specific showing of need for an evaluation by showing that the prosecuting witness may have specific mental or emotional problems directly related to the issues in the case. Second, we require the defendant to demonstrate that the testimony of the person to be examined is uncorroborated or otherwise untrustworthy. Pickens, 675 P.2d at 669. See also Moor v. State, 709 P.2d 498, 508 (Alaska App. 1985).

A number of reasons lead us to conclude that a mental health examination should have been ordered in this case. The state relied heavily on the testimony of Dr. James F. Harper, a clinical and forensic psychologist who specializes in evaluating sexually abused children. Dr. Harper, who did not examine the two children in question, testified to certain behavioral characteristics which he associated with victims of child sexual abuse. J.E.'s and E.E.'s mother then testified that the children exhibited unusual behavior, similar to that testified to by Dr. Harper. This evidence was offered to support an inference that the children had, in fact, been sexually abused.

It appears to us that Dr. Harper's testimony, albeit circumstantially, placed the children's psychological characteristics in controversy. Cf. Schlagenhauf v. Holder, 379 U.S. 104, 117-20, 85 S.Ct. 234, 242-44, 13 L.Ed.2d 152 (1964) (if a litigant's mental health is in controversy in a civil case, the trial court may require an independent mental health examination under the discovery rules). Under the circumstances, a psychological examination of the children would have substantially benefited the defense in preparing to understand and, if possible, to critically evaluate Dr. Harper's testimony. In addition, the children's extreme youth and their answers to certain questions create substantial concern regarding their competency to testify.

Given the broad discretion trial courts have in determining the competency of witnesses, we would not be prepared to say that concerns about competency, standing alone, would warrant psychological evaluations. See, e.g., McMaster v. State, 512 P.2d 879 (Alaska 1973). Nevertheless, when the very real concerns regarding these witnesses' competency are added to the state's substantial reliance on psychological testimony regarding behavioral patterns as a means of identifying sexually abused children, we believe the right to an independent mental health examination has been established. We are particularly concerned that the state apparently used expert testimony to establish that the complaining witnesses and other clients of the day-care center operated by the Andersons were members of a class of sexually-abused children. The foundation offered in support of this testimony was very weak, as is discussed hereafter, and may independently have required reversal.

Finally, we are satisfied that the error in denying the Andersons' request for a mental health examination of the children cannot be termed harmless in this case. As we have noted, Dr. Harper's testimony played a substantial part in the prosecution's case-in-chief. Regardless of whether or not the children would ultimately have been determined competent, an independent psychological examination would certainly have aided the defense in preparing to litigate the question of their competency. We must therefore reverse this case and remand for a new trial. Should the trial court find Dr. Harper's testimony inadmissible, or the state elect to forego it, psychological examinations of the complaining witnesses may be unnecessary.

B. The Grand Jury

Our decision to grant the Andersons a new trial moots a number of the issues they raise. We will nevertheless address some of those issues because they are likely to arise again on retrial. The Andersons have challenged the indictment on a number of grounds. First, they contend that the grand jury foreman, Officer Dennis Long, was employed as a police officer and should not have been permitted to participate in the grand jury proceedings. The Andersons reason that Long, as an employee of the police department that investigates charges of alleged child sexual abuse, too closely identified with the prosecution to be an impartial grand juror.

Even assuming, arguendo, that Long should not have served as a grand juror, we are not prepared to invalidate the indictment because the Andersons have not proved prejudice. They have not established that an insufficient number of grand jurors voted for the indictment so that disqualification of Officer Long would have affected the outcome. Nor have they established that Long was in fact biased, or that he unduly influenced the other grand jurors. They suggest that we should adopt a general rule of disqualification of police officers from grand juries by analogy to their disqualification from petit juries. We are not prepared to adopt such a rule in this case. See Alaska Criminal Rule 6(f)(2) (an indictment shall not be dismissed on the basis that a single grand juror was unqualified if a majority of the qualified grand jurors concurred in finding the indictment).

The Andersons next argue that the prosecutor failed to instruct the grand jury on an essential element of the offense. In the Andersons' view, where sexual penetration is accomplished digitally by someone who has ordinary caretaking responsibilities for children, the grand jury must be instructed that in order to constitute a crime, the penetration must have been for the purpose of gratifying sexual desires. A similar rule has been adopted with regard to sexual contact offenses. See, e.g., Flink v. State, 683 P.2d 725 (Alaska App. 1984). Here, the testimony of the witnesses, if believed, clearly and unequivocally establishes felonious conduct. Looking only at the evidence presented to the grand jury, we find it highly unlikely that the absence of this instruction affected the outcome. We find no error.

The Andersons' next contention is that the prosecutor impermissibly used leading questions in examining the children in connection with the indictment. We find no prejudicial error. See Coger v. State, 517 P.2d 1403 (Alaska 1974).

Finally, the Andersons argue that hearsay was presented to the grand jury without compelling justification, in violation of Alaska Criminal Rule 6(r). Many of the statements about which the Andersons complain are arguably within exceptions to the hearsay rule and therefore were permissibly introduced before the grand jury. If we delete all undisputed hearsay, there was still substantial evidence left supporting the indictment and, consequently, we would not be justified in setting it aside. See, e.g., Frink v. State, 597 P.2d 154, 163-64 (Alaska 1979).

C. Psychological Evidence on Behavior of Abused Children as a Class

With regard to the trial, the Andersons argue that Dr. Harper should not have been permitted to testify. Dr. Harper never examined the children. His testimony was limited to identifying certain behavioral characteristics that mental health professionals associate with victims of sexual abuse. He also testified that children, particularly very young children, are unlikely to fabricate claims of being sexually abused. We have previously expressed concern about expert witnesses purporting to vouch for the credibility of alleged victims of sexual abuse. Colgan v. State, 711 P.2d 533 (Alaska App. 1985).

Lay witnesses testified that J.E. and E.E. manifested some of the behavioral characteristics described by Dr. Harper. In addition, parents of other children who had attended the Andersons' day-care center were permitted to testify that their children also exhibited characteristics that Dr. Harper said are associated with the victims of sexual abuse. The Andersons contend that Dr. Harper's testimony was not within the bounds of proper expert testimony because it did not meet the test established by the District of Columbia Circuit Court of Appeals for "novel scientific evidence." See Frye v. United States, 293 F. 1013 (D.C. Cir. 1923).

We addressed similar issues in Rodriquez v. State, 741 P.2d 1200 (Alaska App. 1987). In Rodriquez, we recognized that several different kinds of expert testimony regarding the behavior of sexually abused children have been offered for a variety of purposes. It is therefore difficult, and perhaps improper, to formulate a single rule to cover all such cases. However, we did isolate one kind of case in which we concluded that such expert testimony is appropriate. When a complaining witness testifies that he or she has been the subject of sexual or physical abuse and the defense seeks to discredit this testimony by showing that the witness' conduct ( i.e., remaining with the allegedly abusive partner or parent or expressing love or affection for that allegedly abusing person) was inconsistent with the claimed abuse and therefore that the claim of abuse was false, the state should be permitted to offer expert testimony that other members of the relevant class ( i.e., abused or battered women or sexually abused children) characteristically exhibit such conduct even though they are, in fact, abused. Compare, D. McCord, Expert Psychological Testimony About Child Complainants in Sexual Abuse Prosecutions: A Foray Into the Admissibility of Novel Psychological Evidence, 77 J.Crim.L. Criminology I (1986) with E.M. Schneider, Describing and Changing: Women's Self-Defense Work and the Problem of Expert Testimony on Battering, 9 Women's Rights Law Reporter 195 (1986).

We have been critical of testimony suggesting that children never lie about sexual abuse. See, e.g., Colgan, 711 P.2d at 535. Moreover, we have never authorized expert testimony seeking to establish that a person is a member of a particular class or group, i.e., battered women or sexually abused children, by showing that they exhibit behavioral characteristics common to that group. We agree with the Andersons that before such testimony is admitted, the proponent should establish, in a hearing out of the presence of the jury, that the probative value of the testimony outweighs its possible prejudicial effect. See, e.g., Alaska Evidence Rule 403. The supreme court has suggested that in cases such as this, compliance with A.R.E. 403 should be virtually the equivalent of compliance with the Frye rule. See, e.g., Contreras v. State, 718 P.2d 129, 136-38 (Alaska 1986).

We stress that in order to determine whether the probative value of such evidence would outweigh its prejudicial effect, the trial court should require the proponent of such evidence to identify in advance specifically what he or she intends to prove and why that evidence would be relevant to the case. In determining the probative value of such expert testimony, the court should certainly consider the extent to which the expert witnesses' assumptions are shared by a consensus of those mental health practitioners knowledgeable about the subject matter. We are particularly concerned in this case because Dr. Harper relied, in part, upon a study reportedly validated by lie detector tests. Given this jurisdiction's rulings regarding the admissibility of polygraph results, Dr. Harper's reliance is problematical. See Pulakis v. State, 476 P.2d 474, 476-80 (Alaska 1970) (expressing substantial reservations about polygraph results).

D. Evidence of Abuse of Other Children

The Andersons also contend that the trial court improperly permitted parents of other children to testify that their children also exhibited the behavioral characteristics identified by Dr. Harper. On remand, if the trial court concludes that Dr. Harper's testimony is inappropriate, this issue will not arise. Should the trial court conclude that Dr. Harper's testimony is admissible, in a form that would provide a foundation for testimony regarding the behavioral characteristics of other children, then the trial court should give careful consideration to whether this testimony complies with Alaska Evidence Rules 404 and 403. Where prior "bad acts" are not validated by a conviction, the trial court should be certain that the evidence of the other bad acts is sufficiently probative to warrant its admission. The state should be required to establish, in a hearing outside the presence of the jury, that each of the other children was, in fact, sexually abused by the Andersons. Mere speculation and conjecture could be particularly prejudicial in a case of this nature. See, e.g., Wright Graham, Federal Practice and Procedure: Evidence § 5249 at 531-37 (1978).

E. Sufficiency of Evidence to Support a Conviction

Finally, the Andersons argue that there was insufficient evidence to support a conviction on Count I of the indictment. They rely on Brower v. State, 728 P.2d 645 (Alaska App. 1986). In our view, this issue is controlled by Bodine v. State, 737 P.2d 1072 (Alaska App. 1987). We find there was sufficient evidence to go to the jury in this case.

The judgment of the superior court is REVERSED and this case REMANDED for trial.