OPINION

SINGLETON, Judge.

Arthur Haakanson was convicted, following a jury trial, of ten counts of a fourteen-count indictment, for sexual abuse of a minor in the first degree, sexual assault in the first degree, and sexual abuse of a minor in the second degree. AS 11.41.434(a)(1); former AS 11.41.410(a)(2); AS 11.41.436(a)(2). Haakanson appeals his conviction and sentence. He first argues that this court should reconsider the admissibility of polygraph examinations and find that the trial court erred in refusing to admit his polygraph examination results. He also contends that the trial court erroneously allowed the state to present testimony describing a "sex offender profile," and that it further erred in admitting physical evidence to fit Haakanson within the profile description. The state concedes that the sex offender profile testimony was improper, but argues it was harmless. Haakanson finally argues that the court erred in allowing the state to introduce evidence of sexual misconduct against children not named in the indictment.

We find that Haakanson has failed to make the necessary showing of reliability to introduce polygraph examination results at trial. We therefore affirm the trial court's order denying admission of Haakanson's polygraph results. Nevertheless, we reverse Haakanson's conviction on the grounds that erroneously admitted character evidence denied him a fair trial.

FACTS AND PROCEEDINGS

Arthur Haakanson was a fifty-five-year-old resident of Old Harbor, a village on Kodiak Island, at the time of trial. The charges against him were based on his alleged sexual contacts with three girls residing in Old Harbor. M.D.C., born November 8, 1972, testified that she had known Haakanson since she was three years old. She claimed that he began touching her breasts and genitals through her clothing when she was in third or fourth grade. S.A.C., born January 14, 1973, testified that he attempted to engage in similar conduct with her. The third victim named in the indictment, B.J.S., born February 26, 1973, testified that she was in third grade when Haakanson first had genital intercourse with her. Over the subsequent two years, she claimed, he engaged in oral, digital, and genital sex with her.

A grand jury issued a fourteen-count indictment against Haakanson on May 3, 1985, for offenses related to his alleged sexual contacts with M.D.C., S.A.C., and B.J.S. On May 30, 1985, Haakanson took a polygraph examination and denied any sexual contact with the three complainants. Haakanson moved to admit his polygraph examination results. Superior Court Judge Rene J. Gonzalez held an evidentiary hearing on Haakanson's motion on September 9 and 10, 1985. The defense presented Richard Slisz, who described the changes in polygraph technology and accuracy in the past fifteen years, how a polygraph test works, and the acceptance among polygraph examiners of the polygraph's reliability to show truthfulness. He claimed that a "well-trained, well-experienced examiner will run in the accuracy rate of eighty-five to ninety-two percent."

Slisz testified that he had asked Haakanson during the polygraph examination if he had touched any of the victims charged in the indictment "with a sexual purpose." His opinion was that Haakanson truthfully denied sexual contact with the victims.

The state then called William Iacono, Associate Professor of Psychology at the University of Minnesota, to testify. Iacono's field of specialty is psychophysiology, which is "the parent science on which [the] polygraph is based." He has written three articles on the polygraph and was a consultant for the United States Office of Technology Assessment's report on the scientific validity of the polygraph.

Iacono testified that the polygraph cannot detect whether a person is "lying" but only whether a person is disturbed when responding to a question. He also testified that studies of the accuracy of polygraphs have yielded varying results and there is presently no consensus of scientific opinion as to the accuracy of polygraph tests.

Judge Gonzalez took Haakanson's motion to admit the polygraph results under advisement. The parties submitted articles and chapters of a book on the accuracy of polygraphs to the court. Judge Gonzalez orally denied the motion prior to trial, and, on February 7, 1986, issued a written order nunc pro tunc September 9, 1985, denying defendant's motion. He found that the "scientific validity of the results of a polygraph examination is not generally accepted in the scientific community sufficiently for its admissibility in a criminal trial."

Jury trial was held in Kodiak in October 1985 with Judge Gonzalez presiding. Alaska State Trooper Rodney Guinn was the state's first witness. He testified that there are certain characteristics which are common to child sexual abusers. Some of the characteristics he mentioned were an unusual interest in children, using treats to attract children, allowing children to do things they may not be allowed to do at home, and paying extra attention to them. He testified that child sexual abusers may own particular items, including sexual aids, sexually explicit photographs, and photos or films of children that are not necessarily sexually explicit.

Defense counsel objected twice to this testimony, based on the ground that Trooper Guinn's testimony described a profile of a "sex offender syndrome." The court overruled the objection, finding that Guinn's testimony was a foundation for the admissibility of some of the state's evidence.

The court subsequently admitted photographs of the three victims, fully clothed, which were found in Haakanson's trailer. Also admitted were a sexually explicit magazine and photo, cut out of the magazine, with the names of M.D.C. and Arthur written above the partially-clad woman and nude man in the photo. The court further admitted bags of candy found in Haakanson's home as well as a game, "Searching for Gold."

The three children named in the indictment testified as previously described. Two other twelve-year-old girls who were not named in the indictment also testified. C.C., age twelve, testified that on two occasions when she visited Haakanson in his home, he held her tight and tried to kiss her on her lips. She stated that he also tried to touch her on her leg, between her knee and pelvis, and that this made her feel "a little bit scared." C.B., age twelve, testified that when she was in Haakanson's trailer, he hugged her a lot, sometimes "real tight" and sometimes "gently." She also corroborated C.C.'s story, testifying that she saw the defendant touch C.C. on the thigh above the knee. The defense concedes that no objection was made to the testimony of C.C. and C.B. at trial.

The trial court also allowed the state, over defense counsel's objection, to ask Haakanson on cross-examination whether he had ever made a "sexually suggestive remark to a child." Haakanson responded that he took several children to a cafe in Old Harbor. While they were there, B.C., a seven-year-old girl, asked him for a quarter and he told her, "If I pull my pants down you'll know where to kiss." On redirect examination, he testified that he was joking and laughing when he made the statement to B.C.

The court then allowed the state to present the testimony of Ernie Berestoff, who worked in the cafe, on rebuttal. Berestoff testified that he heard Haakanson tell B.C., "Wait until I pull down my pants and you know what to kiss," in response to her request for a quarter. The court allowed Berestoff's testimony to rebut Haakanson's assertion that he was joking when he made the statement. Berestoff testified, however, that he could not hear whether Haakanson was laughing at the time he made the statement.

Judge Gonzalez dismissed three of the counts involving S.A.C. at the close of the state's case-in-chief, based on insufficient evidence. The jury returned a guilty verdict on ten of the remaining eleven counts.

ADMISSIBILITY OF HAAKANSON'S POLYGRAPH RESULTS

Haakanson first contends that the trial court erred in finding his polygraph examination results inadmissible at trial. His contention is based on alleged advances in the technology and reliability of polygraphs in the past ten years. He claims that the reliability of polygraphs has been established sufficiently to satisfy the Frye-Pulakis requirements. Frye v. United States, 293 F. 1013 (D.C. Cir. 1923); Pulakis v. State, 476 P.2d 474 (Alaska 1970).

In Pulakis, the Alaska Supreme Court adopted a general rule that polygraph results are inadmissible, based upon the reasoning in Frye. The court did not foreclose the possibility that polygraphs could be shown to be reliable enough to warrant admissibility in the future. Regarding this possibility, the court cautioned:

The central problem regarding admissibility is not that polygraph evidence has proved unreliable, but that polygraph proponents have not yet developed persuasive data demonstrating its reliability. Little worthwhile experimentation has been done to determine the reliability of polygraph evidence. This is not to say that the worth of polygraph evidence cannot ever be proved to the satisfaction of this court. . . . Judicial acceptance of polygraph tests must await the results of more persuasive experimental proof of reliability.

Id. at 479. Polygraph tests have not yet been judicially accepted for use in criminal trials by Alaska appellate courts because they have not been proven reliable. Troyer v. State, 614 P.2d 313, 319 n. 12 (Alaska 1980); Van Meter v. State, 743 P.2d 385, 387-88 (Alaska App. 1987); Leonard v. State, 655 P.2d 766, 770-71 (Alaska App. 1982).

We employ the Frye standards to determine the admissibility of any novel scientific evidence. The Alaska Supreme Court applied Frye in addressing the admissibility of hypnotically refreshed testimony. Contreras v. State, 718 P.2d 129 (Alaska 1986). The court noted that:

Applying Frye is a two-step process: First, the relevant scientific community must be defined, and second, the testimony and publications of the relevant experts in the field must be evaluated to determine if there is general consensus that hypnotically adduced testimony is reliable.

Id. at 135. The court defined the "relevant scientific community" as the "professions which have studied and/or utilized hypnosis for clinical, therapeutic, research and investigative applications." The court specifically excluded "those whose involvement with hypnosis is strictly limited to that of practitioner." Id. Technicians are not considered part of the relevant scientific community because " Frye requires scientific, not merely technical, judgments to be made." Id. It is not the court's duty to determine which experts are correct when applying Frye. The court must instead determine whether there is a consensus among the experts that the challenged technique is reliable. Id. at 135-36. If such a consensus exists, the evidence is admissible. If not, the court must exclude it.

We recently upheld a trial court's refusal to hold an evidentiary hearing on the admissibility of polygraph examinations. Van Meter, 743 P.2d at 388. The defendant, Van Meter, was prepared to call Richard Slisz, who administered his exam, to testify regarding the accuracy of Van Meter's polygraph examination. We found that the offer of proof did not address the requisite broader issue of polygraph testing as a scientific process. Id.

In the present case, Haakanson presented Richard Slisz, apparently the same polygraph examiner, as his sole expert witness at the evidentiary hearing. Slisz is not a member of the "relevant scientific community" for the purposes of determining reliability. Contreras, 718 P.2d at 135. Although he may be a well-qualified and experienced examiner, he falls within the group of technicians, operators, and practitioners which the supreme court excluded in defining the relevant scientific community. Id. Slisz's testimony was therefore insufficient to satisfy the Frye-Contreras standards.

On the other hand, the state's witness, William Iacono, is a psychology professor. He is in a profession which studies the polygraph and its scientific basis. It is fair to surmise that Iacono had a better grasp of the acceptance level of polygraphs within the scientific community than Slisz, inasmuch as he was a member of that community. Iacono appears to be qualified to satisfy the requirements of Frye-Contreras. His testimony, however, does not establish that polygraphs have been accepted as reliable by a consensus of relevant professionals. In fact, he indicated that "no group or society [or] organization has ever been able to reach a consensus of scientific opinion as to exactly how well the test works." The considerable controversy over the reliability of polygraphs as a scientific process is further indicated by the literature submitted to the trial court, as well as the plethora of conflicting literature not introduced. Therefore, Haakanson has not established that there is a consensus among the experts regarding the reliability of the polygraph technique.

We are also concerned about the disproportionate impact polygraph evidence may have on a jury. "[P]olygraph evidence might be perceived by the jury as a complete answer to questions of credibility. Such evidence could also lull the jury into a false sense of security and result in the jury failing to carefully scrutinize conflicting witness testimony." Leonard v. State, 655 P.2d at 770. Any evidence which has such great potential to mislead or prejudice the jury should be excluded unless its probative value clearly outweighs the prejudice. A.R.E. 403. The probative value of polygraph evidence is insubstantial because the polygraph has not been proven reliable. Thus, the prejudicial impact it may have on a jury clearly outweighs the probative value.

The polygraph evidence was inadmissible under Frye-Contreras and under Alaska Evidence Rule 403. Therefore, the trial court did not abuse its discretion in refusing to admit Haakanson's polygraph results at trial. Hawley v. State, 614 P.2d 1349, 1361 (Alaska 1980).

ADMISSIBILITY OF THE "CHILD SEXUAL ABUSER PROFILE"

Haakanson next argues that the trial court erroneously admitted Trooper Guinn's testimony regarding a child sex offender profile. In a related argument, he argues that the admission of evidence fitting him within the profile was prejudicial error.

The state concedes that the admission of Trooper Guinn's testimony regarding a child sexual abuser profile was error because it violated Alaska Rule of Evidence 404(a). When the state concedes error, this court must still make an independent examination of the facts and the law to determine if there was indeed an error. Marks v. State, 496 P.2d 66, 67-68 (Alaska 1972); Roberts v. State, 751 P.2d 507 (Alaska App. 1988).

The Alaska appellate courts have not yet addressed the precise issue of expert testimony on child sexual offender profiles. There is a strong argument that such profile evidence is inadmissible as character evidence under Alaska Evidence Rule 404(a), which reads in part:

Evidence of a person's character or a trait of his character is not admissible for the purpose of proving that he acted in conformity therewith on a particular occasion, except:

(1) Character of Accused. Evidence of a relevant trait of his character offered by an accused, or by the prosecution to rebut the same. . . .

Several jurisdictions have addressed the related issue of a "battering parent" or "child batterer" profile, and have held that evidence of such profiles is inadmissible character evidence. See People v. Walkey, 177 Cal.App.3d 268, 223 Cal. Rptr. 132, 138 (1986) (prosecution may not introduce character evidence of a defendant to show the defendant has the characteristics of a typical battering parent); Sanders v. State, 251 Ga. 70, 303 S.E.2d 13, 18 (1983) (state cannot introduce evidence of a battering parent syndrome unless the defense has first raised the issue); State v. Loebach, 310 N.W.2d 58, 63-64 (Minn. 1981) (prosecution could not introduce evidence of a "battering parent" syndrome or show that the character of the defendant fit the profile of a "battering parent").

At least one jurisdiction has allowed expert testimony explaining pedophilia to the jury. State v. Swallow, 350 N.W.2d 606, 609 (S.D. 1984). The expert in Swallow, however, testified that certain adults are sexually gratified by fondling children or looking at photographs of nude children. This was the same conduct that Swallow was charged with in the indictment. Id. Thus, the prosecution could show that Swallow fit the description of a pedophile by proving that he committed the offenses with which he was charged. The pedophile testimony in this case was substantially different than Trooper Guinn's testimony, which listed otherwise innocent characteristics, such as giving candy to children, as evidence of guilt.

The weight of authority clearly suggests that Rule 404(a) prohibits the profile testimony introduced at trial, unless the defense has raised the issue first. See E. Cleary, McCormick on Evidence § 206(D) (3d ed. 1984 Supp. 1987). We hold that the prosecution may not introduce a profile to show that the defendant is more likely to have committed an offense because the defendant fits within that profile. To admit this testimony at the beginning of trial was clearly erroneous.

We are also concerned that a jury may place undue emphasis on sex offender profile testimony. Alaska Evidence Rule 403 provides for excluding relevant evidence if its "probative value is outweighed by the danger of unfair prejudice, confusion of the issues, or misleading the jury[.]" Other jurisdictions have found profile evidence introduced to establish that the defendant is a member of a particular group inadmissible on the basis that the prejudicial danger of such evidence outweighs any probative value it may have. The Maryland Court of Appeals, found that "[child batterer] profile evidence is highly prejudicial because it invites a jury to conclude that because an expert experienced in child abuse cases identifies an accused as someone fitting a particular profile, it is more likely than not that this individual committed the crime." Sloan v. State, 70 Md. App. 630, 522 A.2d 1364, 1368 (1987). Washington appellate courts have also held that evidence which identifies the defendant as a member of a group with a statistically higher incidence of child sexual abuse is inadmissible because of the high danger of prejudice. State v. Petrich, 101 Wn.2d 566, 683 P.2d 173, 180 (1984); State v. Claflin, 690 P.2d 1186, 1190 (Wash.App. 1984); State v. Maule, 35 Wn. App. 287, 667 P.2d 96, 99 (1983).

We have recently addressed the related issue of the admissibility of expert testimony regarding behavioral characteristics common to child victims of sexual abuse. Anderson v. State, 749 P.2d 369, 373-74 (Alaska App. 1988); Rodriquez v. State, 741 P.2d 1200, 1203-05 (Alaska App. 1987). Rodriquez and Anderson, read together, permit expert testimony that responds to a defense claim that a complaining witness' conduct is inconsistent with being sexually abused by showing that similar conduct is exhibited by those who are sexually abused. These decisions do not permit testimony offered to prove that the complaining witness is sexually abused by showing that the complaining witness exhibits behavior similar to that exhibited by sexually abused children.

We stated in Anderson:

[W]e 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 . . . 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 prejudicial effect. The [Alaska] [S]upreme [C]ourt has suggested that in cases such as this, compliance with [Evidence Rule] 403 should be virtually the equivalent of compliance with the Frye rule.

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.

749 P.2d 373-74 (citations omitted).

The state did not establish that Guinn's assumptions were shared by a consensus of experts in the mental health field nor that Guinn was a member of the relevant group of experts. See Contreras v. State, 718 P.2d at 135. Even if the profile evidence did not offend the principles embodied in Evidence Rule 404(a), it would not be admissible under the standard articulated in Anderson. The state did not, and probably could not, show that the probative value of Guinn's testimony outweighed the inherent prejudicial effect of the profile evidence. Therefore, the trial court erred in allowing Trooper Guinn to testify regarding the characteristics of a typical child sexual abuser because such profile testimony was inadmissible under Evidence Rules 403 and 404(a).

The state claims that the admission of Guinn's testimony regarding the profile was harmless error. To support its argument, it notes that the profile was not mentioned by the prosecution in the opening statement or closing argument. It also points out that the state's case against Haakanson was strong; three separate victims testified to sexual abuse by Haakanson.

Haakanson counters that the erroneous admission of the "profile" testimony was not harmless. He claims that the "state spun a web of evidence showing the jury how the defendant just happened to have many of the same attributes of a child molester." He also contends that the timing of the profile testimony at the commencement of trial focused the jury's attention from the start on whether he had the characteristics of a child molester. Because of the prejudicial impact of the timing, he posits, the fact that the prosecutor did not mention the profile in closing arguments does not render the error harmless. He also contends that the state's case was not that strong without the profile testimony and the evidence comparing the defendant to that profile.

Whether erroneously admitted evidence justifies a reversal depends on whether the evidence "would have had a substantial effect upon a jury's deliberations." Clifton v. State, 751 P.2d 27, 30 (Alaska 1988) (quoting Spaulding v. State, 481 P.2d 389, 393 (Alaska 1971)). We recognize that other jurisdictions have found the admission of profile testimony to be harmless when the state's case is otherwise strong. Walkey, 223 Cal.Rptr. at 139; Loebach, 310 N.W.2d at 64. In this case, however, the profile evidence tainted most of the state's case. The profile provided a foundation for admission of many items of evidence, including a pornographic magazine. The presentation of the profile evidence at the beginning of trial provided the jury with an initial framework within which to judge Haakanson's actions. Such profile evidence would likely have a substantial effect on jury deliberations. Clifton, 751 P.2d at 30. We hold that this evidence was sufficiently prejudicial to warrant reversal. Id., Love v. State, 457 P.2d 622 (Alaska 1969).

ADMISSIBILITY OF PRIOR ACTS

Haakanson claims that the trial court erroneously admitted evidence of his statement to B.C. in the cafe and of his conduct with C.C. and C.B. He contends that Alaska Evidence Rule 404(b) prohibits the state's use of his prior acts as evidence of his propensity to commit the crimes charged in the indictment.

Evidence of sexual conduct involving victims not named in the indictment is generally inadmissible under Evidence Rule 404(b). Bolden v. State, 720 P.2d 957 (Alaska App. 1986). In addition, the court must consider Evidence Rule 403 in conjunction with Rule 404(b). Bolden v. State, 720 P.2d at 959; Lerchenstein v. State, 697 P.2d 312, 315-16 (Alaska App. 1985), aff'd, 726 P.2d 546 (Alaska 1986).

In cases of sexual abuse, the trial court must "carefully scrutinize" evidence of uncharged sexual acts. Bolden, 720 P.2d at 960; Moor v. State, 709 P.2d 498, 506 (Alaska App. 1985). When evidence of prior bad acts is offered, the court first must determine that the evidence sought to be admitted has relevance apart from propensity. If there is nonpropensity relevance, the court must then determine whether that relevance outweighs the presumed highly prejudicial impact of the evidence. Bolden, 720 P.2d at 960 (quoting Lerchenstein, 697 P.2d at 315-16).

The state contends that Berestoff's testimony regarding Haakanson's statement to B.C. in the cafe was relevant to show Haakanson's state of mind. The trial court overruled the defense objection to this line of inquiry during cross-examination of Haakanson, reasoning that Haakanson's statement to B.C. was relevant to his state of mind towards the victims named in the indictment. The court further found that its probative value as evidence of Haakanson's state of mind outweighed any prejudicial effect it might have under Rule 403.

We have previously noted that "[e]vidence regarding a motive to have intercourse with one woman does not support an inference that the defendant was motivated to have intercourse with a second." Pletnikoff v. State, 719 P.2d 1039, 1043 (Alaska App. 1986). There was no nexus between Haakanson's statement to B.C. and his sexual intentions towards the children named in the indictment. We reject the contention that this evidence was relevant to his state of mind regarding the offenses. Because the evidence had no nonpropensity value, the court did not need to reach the balancing process under Rule 403. Bolden, 720 P.2d at 959.

The court further erred in allowing the state to question Berestoff about this incident. Berestoff's testimony, purported to rebut Haakanson's claim that he was joking with B.C., is also inadmissible because it sought to impeach Haakanson on a collateral matter. Moss v. State, 620 P.2d 674 (Alaska 1980). The state contends that any error was harmless. See Alaska R.Crim.P. 47(a). In light of our holding that the admission of profile evidence justified a new trial, and the recent amendments to Rule 404(b), we do not decide here whether the error warranted reversal.

The testimony of C.C. and C.B. regarding Haakanson's inappropriate physical behavior with them was also inadmissible under Evidence Rules 403 and 404(b). Haakanson did not, however, object to their testimony at trial. Haakanson contends that although there was no objection, the admission of their testimony amounts to plain error. Alaska R.Crim.P. 47(b). Our decision to reverse on other grounds, however, makes it unnecessary to decide whether admission of these acts constituted plain error under Rule 47(b).

The conviction is REVERSED and the case REMANDED for a new trial.